We address the problem of estimating both translational and angular velocity of a camera from asynchronous point tracks, a formulation relevant to rolling shutter and event cameras. Since the original problem is non-polynomial, we propose a polynomial approximation, classify the resulting minimal problems, and determine their algebraic degrees. Furthermore, we develop minimal solvers for several problems with low degrees and evaluate them on synthetic and real datasets. The code is publicly available.

This paper introduces RaCo, a lightweight neural network designed to learn robust and versatile keypoints suitable for a variety of 3D computer vision tasks. The model integrates three key components: the repeatable keypoint detector, a differentiable ranker to maximize matches with a limited number of keypoints, and a covariance estimator to quantify spatial uncertainty in metric scale. Trained on perspective image crops only, RaCo operates without the need for covisible image pairs. It achieves strong rotational robustness through extensive data augmentation, even without the use of computationally expensive equivariant network architectures. The method is evaluated on several challenging datasets, where it demonstrates state-of-the-art performance in keypoint repeatability and two-view matching, particularly under large in-plane rotations. Ultimately, RaCo provides an effective and simple strategy to independently estimate keypoint ranking and metric covariance without additional labels, detecting interpretable and repeatable interest points. The code is available at: https://github.com/cvg/RaCo

Abstract Visual odometry (VO) is fundamental to autonomous navigation, robotics, and augmented reality. While self-supervised learning has eliminated the need for expensive ground-truth labels in monocular VO, dynamic objects and occlusions that violate the static scene assumption lead to erroneous pose estimates. Existing uncertainty-based methods filter unreliable regions but rely solely on single-frame information, neglecting temporal consistency across consecutive frames. We present Combined Projected Uncertainty (CoProU), a principled probabilistic formulation that propagates and fuses uncertainties across temporal frames. Our key insight is that robust uncertainty estimation requires combining target frame uncertainty with projected uncertainty from reference frames, enabling effective identification of dynamic regions and temporal inconsistencies. We demonstrate CoProU’s versatility through two complementary frameworks. CoProU-VO-2F employs a decoupled architecture with CNN-based pose encoder and vision transformer-based depth encoder for two-frame visual odometry. CoProU-VO-MF extends our approach to multi-frame scenarios using a unified transformer architecture with coupled encoders that produce shared representations for ego-motion and geometry estimation. This demonstrates that CoProU, though originally formulated for frame pairs, generalizes naturally to multi-frame settings through pairwise application. Comprehensive experiments validate our contributions. CoProU-VO-2F achieves substantial improvements over state-of-the-art two-frame methods, reducing ATE by up to 63% on KITTI and 33% on nuScenes. CoProU-VO-MF achieves 45% lower average ATE across KITTI, nuScenes, and Waymo compared to the large-scale pretrained VGGT baseline. Extensive ablation studies confirm the effectiveness of temporal uncertainty propagation and CoProU’s adaptability across different architectural paradigms. Please check out our Project Page .

3D reconstruction aims to recover the dense 3D structure of a scene. It plays an essential role in various applications such as Augmented/Virtual Reality (AR/VR), autonomous driving and robotics. Leveraging multiple views of a scene captured from different viewpoints, Multi-View Stereo (MVS) algorithms synthesize a comprehensive 3D representation, enabling precise reconstruction in complex environments. Due to its efficiency and effectiveness, MVS has become a pivotal method for image-based 3D reconstruction. Recently, with the success of deep learning, many learning-based MVS methods have been proposed, achieving impressive performance against traditional methods. We categorize these learning-based methods as: depth map-based, voxel-based, NeRF-based, 3D Gaussian Splatting-based, and large feed-forward methods. Among these, we focus significantly on depth map-based methods, which are the main family of MVS due to their conciseness, flexibility and scalability. In this survey, we provide a comprehensive review of the literature at the time of this writing. We investigate these learning-based methods, summarize their performances on popular benchmarks, and discuss promising future research directions in this area.

We propose SNI-SLAM++, a tightly-coupled semantic SLAM system utilizing neural implicit representation, that simultaneously performs accurate semantic mapping, high-quality surface reconstruction, and robust camera tracking. Our system tightly integrates visual appearance, geometry, and semantics through five key components: (i) We introduce hierarchical semantic representation to allow multi-level semantic comprehension for top-down structured semantic mapping of the scene. (ii) To fully utilize the correlation between multiple attributes of the environment, we integrate appearance, geometry and semantic features through cross-attention for feature collaboration. This strategy enables a more multifaceted understanding of the environment, thereby allowing SNI-SLAM++ to remain robust even when single attribute is defective. (iii) We design an internal fusion-based decoder to obtain semantic, RGB, and Truncated Signed Distance Field (TSDF) values from multi-level features for accurate decoding. (iv) We introduce a semantics-coupled tracking framework that tightly incorporates semantic constraints for camera pose estimation in neural implicit SLAM. This framework leverages the multi-view consistency of semantics to construct a pose graph and perform semantic loop closure optimization, enabling robust tracking. (v) We propose a feature loss to update the scene representation at the feature level. Compared with low-level losses such as RGB loss and depth loss, our feature loss is capable of guiding the network optimization on a higher level. Our SNI-SLAM++ demonstrates superior performance over all recent visual SLAM methods in terms of mapping and tracking accuracy on the datasets of Replica, ScanNet, TUM-RGBD, and ScanNet++, while also showing excellent capabilities in accurate semantic segmentation and 3D semantic mapping.

Large diffusion models demonstrate remarkable zeroshot capabilities in novel view synthesis from a single image. However, these models often face challenges in maintaining consistency across novel and reference views. A crucial factor leading to this issue is the limited utilization of contextual information from reference views. Specifically, when there is an overlap in the viewing frustum between two views, it is essential to ensure that the corresponding regions maintain consistency in both geometry and appearance. This observation leads to a simple yet effective approach, where we propose to use epipolar geometry to locate and retrieve overlapping information from the input view. This information is then incorporated into the generation of target views, eliminating the need for training or fine-tuning, as the process requires no learnable parameters. Furthermore, to enhance the overall consistency of generated views, we extend the utilization of epipolar attention to a multi-view setting, allowing retrieval of overlapping information from the input view and other target views. Qualitative and quantitative experimental results demonstrate the effectiveness of our method in significantly improving the consistency of synthesized views without the need for any fine-tuning. Moreover, This enhancement also boosts the performance of downstream applications such as 3D reconstruction. The code is available at https://github.com/botaoye/ConsisSyn.

Human activities are inherently complex, and even simple household tasks involve numerous object interactions. To better understand these activities, it is crucial to model their interactions with the environment captured through dynamic changes. The recent availability of affordable head-mounted cameras and egocentric data offers a more accessible and efficient means to understand dynamic human-object interactions in 3D environments. However, most existing methods for human activity modeling either focus on reconstructing 3D models of hand-object or human-scene interactions or on mapping 3D scenes, neglecting dynamic interactions with objects. The few existing solutions often require inputs from multiple sources, including multi-camera setups, depth-sensing cameras, or kinesthetic sensors. To this end, we introduce EgoGaussian, the first method capable of simultaneously reconstructing 3D scenes and dynamically tracking 3D object motion from RGB egocentric input alone. We leverage the uniquely discrete nature of Gaussian Splatting and segment dynamic interactions from the background. Our approach employs a clip-level online learning pipeline that leverages the dynamic nature of human activities, allowing us to reconstruct the temporal evolution of the scene in chronological order and track rigid object motion. Additionally, our method automatically segments object and background Gaussians, providing explicit 3D representations for both static scenes and dynamic objects. EgoGaussian shows significant improvements in terms of both dynamic object and background reconstruction quality compared to the state-of-the-art. We also qualitatively demonstrate the high quality of the reconstructed models.

Monocular depth estimation (MDE) models have undergone significant advancements over recent years. Many MDE models aim to predict affine-invariant relative depth from monocular images, while recent developments in large-scale training and vision foundation models enable reasonable estimation of metric (absolute) depth. However, effectively leveraging these predictions for geometric vision tasks, in particular relative pose estimation, remains relatively under explored. While depths provide rich constraints for cross-view image alignment, the intrinsic noise and ambiguity from the monocular depth priors present practical challenges to improving upon classic keypoint-based solutions. In this paper, we develop three solvers for relative pose estimation that explicitly account for independent affine (scale and shift) ambiguities, covering both calibrated and uncalibrated conditions. We further propose a hybrid estimation pipeline that combines our proposed solvers with classic point-based solvers and epipolar constraints. We find that the affine correction modeling is beneficial to not only the relative depth priors but also, surprisingly, the "metric" ones. Results across multiple datasets demonstrate large improvements of our approach over classic keypoint-based baselines and PnP-based solutions, under both calibrated and uncalibrated setups. We also show that our method improves consistently with different feature matchers and MDE models, and can further benefit from very recent advances on both modules. Code is available at https://github.com/MarkYu98/madpose.

Accurate localization plays a pivotal role in the autonomy of systems operating in unfamiliar environments, particularly when interaction with humans is expected. High-accuracy visual localization systems encompass various components, such as image retrievers, feature extractors, matchers, reconstruction and pose estimation methods. This complexity translates to the necessity of robust evaluation settings and pipelines. However, existing datasets and benchmarks primarily focus on single-agent scenarios, overlooking the critical issue of cross-device localization. Different agents with different sensors will show their own specific strengths and weaknesses, and the data they have available varies substantially. This work addresses this gap by enhancing an existing augmented reality visual localization benchmark with data from legged robots, and evaluating human-robot, cross-device mapping and localization. Our contributions extend beyond device diversity and include high environment variability, spanning ten distinct locations ranging from disaster sites to art exhibitions. Each scene in our dataset features recordings from robot agents, hand-held and head-mounted devices, and high-accuracy ground truth LiDAR scanners, resulting in a comprehensive multi-agent dataset and benchmark. This work represents a significant advancement in the field of visual localization benchmarking, with key in-sights into the performance of cross-device localization methods across diverse settings.

Future robots are envisioned as versatile systems capable of performing a variety of household tasks. The big question remains, how can we bridge the embodiment gap while minimizing physical robot learning, which fundamentally does not scale well. We argue that learning from in-the-wild human videos offers a promising solution for robotic manipulation tasks, as vast amounts of relevant data already exist on the internet. In this work, we present VidBot, a framework enabling zero-shot robotic manipulation using learned 3D affordance from in-the-wild monocular RGB-only human videos. VidBot leverages a pipeline to extract explicit representations from them, namely 3D hand trajectories from videos, combining a depth foundation model with structure-from-motion techniques to reconstruct temporally consistent, metric-scale 3D affordance representations agnostic to embodiments. We introduce a coarse-to-fine affordance learning model that first identifies coarse actions from the pixel space and then generates fine-grained interaction trajectories with a diffusion model, conditioned on coarse actions and guided by test-time constraints for context-aware interaction planning, enabling substantial generalization to novel scenes and embodiments. Extensive experiments demonstrate the efficacy of VidBot, which significantly outperforms counterparts across 13 manipulation tasks in zero-shot settings and can be seamlessly deployed across robot systems in real-world environments. VidBot paves the way for leveraging everyday human videos to make robot learning more scalable.

We present a novel approach to enhance camera pose estimation in global Structure-from-Motion (SfM) frameworks by filtering inaccurate pose graph edges – representing relative translation estimates – before applying translation averaging. In SfM, pose graph vertices represent images, and edges represent relative poses (rotations and translations) between cameras. We reformulate the edge filtering problem as a vertex filtering in the dual graph, specifically, a line graph where vertices correspond to edges in the original graph and edges correspond to cameras. Utilizing this representation, we frame the problem as a binary classification over nodes in the dual graph. To identify outlier edges, we employ a Transformer-based architecture. To overcome the challenge of memory overflow caused by converting to a line graph, we introduce a clustering-based graph processing approach, enabling our method to be applied to arbitrarily large pose graphs. Our method outperforms existing relative translation filtering techniques in terms of camera position accuracy and can be seamlessly integrated with other filters. The code is available at https://github.com/DmblnNicole/LFOE-GlobalSfM.

We present GEM, a Generalizable Ego-vision Multimodal world model that predicts future frames using a reference frame, sparse features, human poses, and ego-trajectories. Hence, our model has precise control over object dynamics, ego-agent motion and human poses. GEM generates paired RGB and depth outputs for richer spatial understanding. We introduce autoregressive noise schedules to enable stable long-horizon generations. Our dataset is comprised of 4000+ hours of multimodal data across domains like autonomous driving, egocentric human activities, and drone flights. Pseudo-labels are used to get depth maps, ego-trajectories, and human poses. We use a comprehensive evaluation framework, including a new Control of Object Manipulation (COM) metric, to assess controllability. Experiments show GEM excels at generating diverse, controllable scenarios and temporal consistency over long generations. Code, models, and datasets are fully open-sourced1.

Neural network performance scales with both model size and data volume, as shown in both language and image processing. This requires scaling-friendly architectures and large datasets. While transformers have been adapted for 3D vision, a ‘GPT-moment’ remains elusive due to limited training data. We introduce ARKit LabelMaker, a large-scale real-world 3D dataset with dense semantic annotation that is more than three times larger than prior largest dataset. Specifically, we extend ARKitScenes [4] with automatically generated dense 3D labels using an extended LabelMaker pipeline [37], tailored for large-scale pre-training. Training on our dataset improves accuracy across architectures, achieving state-of-the-art 3D semantic segmentation scores on ScanNet and ScanNet200, with notable gains on tail classes. Our code is available at labelmaker.org and our dataset at huggingface.

Learning-based visual localization methods that use scene coordinate regression (SCR) offer the advantage of smaller map sizes. However, on datasets with complex illumination changes or image-level ambiguities, it remains a less robust alternative to feature matching methods. This work aims to close the gap. We introduce a covisibility graph-based global encoding learning and data augmentation strategy, along with a depth-adjusted reprojection loss to facilitate implicit triangulation. Additionally, we revisit the network architecture and local feature extraction module. Our method achieves state-of-the-art on challenging large-scale datasets without relying on network ensembles or 3D supervision. On Aachen Day-Night, we are 10× more accurate than previous SCR methods with similar map sizes and require at least 5× smaller map sizes than any other SCR method while still delivering superior accuracy. Code is available at: https://github.com/cvg/scrstudio.

While Structure-From-Motion (SfM) has seen much progress over the years, state-of-the-art systems are prone to failure when facing extreme viewpoint changes in low-overlap, low-parallax or high-symmetry scenarios. Because capturing images that avoid these pitfalls is challenging, this severely limits the wider use of SfM, especially by non-expert users. We overcome these limitations by augmenting the classical SfM paradigm with monocular depth and normal priors inferred by deep neural networks. Thanks to a tight integration of monocular and multi-view constraints, our approach significantly outperforms existing ones under extreme viewpoint changes, while maintaining strong performance in standard conditions. We also show that monocular priors can help reject faulty associations due to symmetries, which is a long-standing problem for SfM. This makes our approach the first capable of reliably reconstructing challenging indoor environments from few images. Through principled uncertainty propagation, it is robust to errors in the priors, can handle priors inferred by different models with little tuning, and will thus easily benefit from future progress in monocular depth and normal estimation. Our code is publicly available at github.com/cvg/mpsfm.

Multi-modal 3D object understanding has gained significant attention, yet current approaches often assume complete data availability and rigid alignment across all modalities. We present CrossOver, a novel framework for cross-modal 3D scene understanding via flexible, scene-level modality alignment. Unlike traditional methods that require aligned modality data for every object instance, CrossOver learns a unified, modality-agnostic embedding space for scenes by aligning modalities – RGB images, point clouds, CAD models, floorplans, and text descriptions – with relaxed constraints and without explicit object semantics. Leveraging dimensionality-specific encoders, a multi-stage training pipeline, and emergent cross-modal behaviors, CrossOver supports robust scene retrieval and object localization, even with missing modalities. Evaluations on Scan-Net and 3RScan datasets show its superior performance across diverse metrics, highlighting CrossOver’s adaptability for real-world applications in 3D scene understanding.

In this paper, we present a new generic Structure-From-Motion pipeline, GenSfM, that uses a non-parametric camera projection model. The model is self-calibrated during the reconstruction process and can fit a wide variety of cameras, ranging from simple low-distortion pinhole cameras to more extreme optical systems such as fisheye or catadioptric cameras. The key component in our framework is an adaptive calibration procedure that can estimate partial calibrations, only modeling regions of the image where sufficient constraints are available. In experiments, we show that our method achieves comparable accuracy to traditional Structure-From-Motion pipelines in easy scenarios, and outperforms them in cases where they are unable to self-calibrate their parametric models. Code is at https://github.com/Ivonne320/GenSfM.git

Gaussian splatting and single-view depth estimation are typically studied in isolation. In this paper, we present Depth-Splat to connect Gaussian splatting and depth estimation and study their interactions. More specifically, we first contribute a robust multi-view depth model by leveraging pretrained monocular depth features, leading to high-quality feed-forward 3D Gaussian splatting reconstructions. We also show that Gaussian splatting can serve as an unsupervised pre-training objective for learning powerful depth models from large-scale multi-view posed datasets. We validate the synergy between Gaussian splatting and depth estimation through extensive ablation and cross-task transfer experiments. Our DepthSplat achieves state-of-the-art performance on ScanNet, RealEstate10K and DL3DV datasets in terms of both depth estimation and novel view synthesis, demonstrating the mutual benefits of connecting both tasks. In addition, DepthSplat enables feed-forward reconstruction from 12 input views (512 × 960 resolutions) in 0.6 seconds.

We introduce the task of predicting functional 3D scene graphs for real-world indoor environments from posed RGB-D images. Unlike traditional 3D scene graphs that focus on spatial relationships of objects, functional 3D scene graphs capture objects, interactive elements, and their functional relationships. Due to the lack of training data, we leverage foundation models, including visual language models (VLMs) and large language models (LLMs), to encode functional knowledge. We evaluate our approach on an extended SceneFun3D dataset and a newly collected dataset, FunGraph3D, both annotated with functional 3D scene graphs. Our method significantly outperforms adapted baselines, including Open3DSG and ConceptGraph, demonstrating its effectiveness in modeling complex scene functionalities. We also demonstrate downstream applications such as 3D question answering and robotic manipulation using functional 3D scene graphs. See our project page at https://openfungraph.github.io.

Touch contact and pressure are essential for understanding how humans interact with objects and offer insights that benefit applications in mixed reality and robotics. Estimating these interactions from an egocentric camera perspective is challenging, largely due to the lack of comprehensive datasets that provide both hand poses and pressure annotations. In this paper, we present EgoPressure, an egocentric dataset that is annotated with high-resolution pressure intensities at contact points and precise hand pose meshes, obtained via our multi-view, sequence-based optimization method. We introduce baseline models for estimating applied pressure on external surfaces from RGB images, both with and without hand pose information, as well as a joint model for predicting hand pose and the pressure distribution across the hand mesh. Our experiments show that pressure and hand pose complement each other in understanding hand-object interactions.

We present WildGS-SLAM, a robust and efficient monocular RGB SLAM system designed to handle dynamic environments by leveraging uncertainty-aware geometric mapping. Unlike traditional SLAM systems, which assume static scenes, our approach integrates depth and uncertainty information to enhance tracking, mapping, and rendering performance in the presence of moving objects. We introduce an uncertainty map, predicted by a shallow multi-layer perceptron and DI-NOv2 features, to guide dynamic object removal during both tracking and mapping. This uncertainty map enhances dense bundle adjustment and Gaussian map optimization, improving reconstruction accuracy. Our system is evaluated on multiple datasets and demonstrates artifact-free view synthesis. Results showcase WildGS-SLAM’s superior performance in dynamic environments compared to state-of-the-art methods.

In dynamic 3D environments, accurately updating scene representations over time is crucial for applications in robotics, mixed reality, and embodied AI. As scenes evolve, efficient methods to incorporate changes are needed to maintain up-to-date, high-quality reconstructions without the computational overhead of re-optimizing the entire scene. This paper introduces CL-Splats, which incrementally updates Gaussian splatting-based 3D representations from sparse scene captures. CL-Splats integrates a robust change-detection module that segments updated and static components within the scene, enabling focused, local optimization that avoids unnecessary re-computation. Moreover, CL-Splats supports storing and recovering previous scene states, facilitating temporal segmentation and new scene-analysis applications. Our extensive experiments demonstrate that CL-Splats achieves efficient updates with improved reconstruction quality over the state-of-the-art. This establishes a robust foundation for future real-time adaptation in 3D scene reconstruction tasks.

We introduce HouseTour, a method for spatially-aware 3D camera trajectory and natural language summary generation from a collection of images depicting an existing 3D space. Unlike existing vision-language models (VLMs), which struggle with geometric reasoning, our approach generates smooth video trajectories via a diffusion process constrained by known camera poses and integrates this information into the VLM for 3D-grounded descriptions. We synthesize the final video using 3D Gaussian splatting to render novel views along the trajectory. To support this task, we present the HouseTour dataset, which includes over 1,200 house-tour videos with camera poses, 3D reconstructions, and real estate descriptions. Experiments demonstrate that incorporating 3D camera trajectories into the text generation process improves performance over methods handling each task independently. We evaluate both individual and end-to-end performance, introducing a new joint metric. Our work enables automated, professional-quality video creation for real estate and touristic applications without requiring specialized expertise or equipment.

We present SuperDec, an approach for creating compact 3D scene representations via decomposition into superquadric primitives. While most recent works leverage geometric primitives to obtain photorealistic 3D scene representations, we propose to leverage them to obtain a compact yet expressive representation. We propose to solve the problem locally on individual objects and leverage the capabilities of instance segmentation methods to scale our solution to full 3D scenes. In doing that, we design a new architecture which efficiently decompose point clouds of arbitrary objects in a compact set of superquadrics. We train our architecture on ShapeNet and we prove its generalization capabilities on object instances extracted from the ScanNet++ dataset as well as on full Replica scenes. Finally, we show how a compact representation based on superquadrics can be useful for a diverse range of downstream applications, including robotic tasks and controllable visual content generation and editing.

We propose a novel approach for estimating the relative pose between rolling shutter cameras using the intersections of line projections with a single scanline per image. This allows pose estimation without explicitly modeling camera motion. Alternatively, scanlines can be selected within a single image, enabling single-view relative pose estimation for scanlines of rolling shutter cameras. Our approach is designed as a foundational building block for rolling shutter structure-from-motion (SfM), where no motion model is required, and each scanline's pose can be computed independently. % We classify minimal solvers for this problem in both generic and specialized settings, including cases with parallel lines and known gravity direction, assuming known intrinsics and no lens distortion. Furthermore, we develop minimal solvers for the parallel-lines scenario, both with and without gravity priors, by leveraging connections between this problem and the estimation of 2D structure from 1D cameras. % Experiments on rolling shutter images from the Fastec dataset demonstrate the feasibility of our approach for initializing rolling shutter SfM, highlighting its potential for further development. % The code will be made publicly available.

We introduce the first data-driven multi-view 3D point tracker, designed to track arbitrary points in dynamic scenes using multiple camera views. Unlike existing monocular trackers, which struggle with depth ambiguities and occlusion, or prior multi-camera methods that require over 20 cameras and tedious per-sequence optimization, our feed-forward model directly predicts 3D correspondences using a practical number of cameras (e.g., four), enabling robust and accurate online tracking. Given known camera poses and either sensor-based or estimated multi-view depth, our tracker fuses multi-view features into a unified point cloud and applies k-nearest-neighbors correlation alongside a transformer-based update to reliably estimate long-range 3D correspondences, even under occlusion. We train on 5K synthetic multi-view Kubric sequences and evaluate on two real-world benchmarks: Panoptic Studio and DexYCB, achieving median trajectory errors of 3.1 cm and 2.0 cm, respectively. Our method generalizes well to diverse camera setups of 1-8 views with varying vantage points and video lengths of 24-150 frames. By releasing our tracker alongside training and evaluation datasets, we aim to set a new standard for multi-view 3D tracking research and provide a practical tool for real-world applications. Project page available at https://ethz-vlg.github.io/mvtracker.

Monocular 3D object detection is valuable for various applications such as robotics and AR/VR. Existing methods are confined to closed-set settings, where the training and testing sets consist of the same scenes and/or object categories. However, real-world applications often introduce new environments and novel object categories, posing a challenge to these methods. In this paper, we address monocular 3D object detection in an open-set setting and introduce the first end-to-end 3D Monocular Open-set Object Detector (3D-MOOD). We propose to lift the open-set 2D detection into 3D space through our designed 3D bounding box head, enabling end-to-end joint training for both 2D and 3D tasks to yield better overall performance. We condition the object queries with geometry prior and overcome the generalization for 3D estimation across diverse scenes. To further improve performance, we design the canonical image space for more efficient cross-dataset training. We evaluate 3D-MOOD on both closed-set settings (Omni3D) and open-set settings (Omni3D to Argoverse 2, ScanNet), and achieve new state-of-the-art results. Code and models are available at royyang0714.github.io/3D-MOOD.

We introduce NoPoSplat, a feed-forward model capable of reconstructing 3D scenes parameterized by 3D Gaussians from unposed sparse multi-view images. Our model, trained exclusively with photometric loss, achieves real-time 3D Gaussian reconstruction during inference. To eliminate the need for accurate pose input during reconstruction, we anchor one input view's local camera coordinates as the canonical space and train the network to predict Gaussian primitives for all views within this space. This approach obviates the need to transform Gaussian primitives from local coordinates into a global coordinate system, thus avoiding errors associated with per-frame Gaussians and pose estimation. To resolve scale ambiguity, we design and compare various intrinsic embedding methods, ultimately opting to convert camera intrinsics into a token embedding and concatenate it with image tokens as input to the model, enabling accurate scene scale prediction. We utilize the reconstructed 3D Gaussians for novel view synthesis and pose estimation tasks and propose a two-stage coarse-to-fine pipeline for accurate pose estimation. Experimental results demonstrate that our pose-free approach can achieve superior novel view synthesis quality compared to pose-required methods, particularly in scenarios with limited input image overlap. For pose estimation, our method, trained without ground truth depth or explicit matching loss, significantly outperforms the state-of-the-art methods with substantial improvements. This work makes significant advances in pose-free generalizable 3D reconstruction and demonstrates its applicability to real-world scenarios. Code and trained models are available at https://noposplat.github.io/.

Understanding behavior requires datasets that capture humans while carrying out complex tasks. The kitchen is an excellent environment for assessing human motor and cognitive function, as many complex actions are naturally exhibited in kitchens from chopping to cleaning. Here, we introduce the EPFL-Smart-Kitchen-30 dataset, collected in a noninvasive motion capture platform inside a kitchen environment. Nine static RGB-D cameras, inertial measurement units (IMUs) and one head-mounted HoloLens~2 headset were used to capture 3D hand, body, and eye movements. The EPFL-Smart-Kitchen-30 dataset is a multi-view action dataset with synchronized exocentric, egocentric, depth, IMUs, eye gaze, body and hand kinematics spanning 29.7 hours of 16 subjects cooking four different recipes. Action sequences were densely annotated with 33.78 action segments per minute. Leveraging this multi-modal dataset, we propose four benchmarks to advance behavior understanding and modeling through 1) a vision-language benchmark, 2) a semantic text-to-motion generation benchmark, 3) a multi-modal action recognition benchmark, 4) a pose-based action segmentation benchmark. We expect the EPFL-Smart-Kitchen-30 dataset to pave the way for better methods as well as insights to understand the nature of ecologically-valid human behavior. Code and data are available at https://amathislab.github.io/EPFL-Smart-Kitchen

Automating the expert-dependent and labor-intensive task of 3D scene synthesis would significantly benefit fields such as architectural design, robotics simulation, and virtual reality. Recent approaches to 3D scene synthesis often rely on the commonsense reasoning of large language models (LLMs) or strong visual priors from image generation models. However, current LLMs exhibit limited 3D spatial reasoning, undermining the realism and global coherence of synthesized scenes, while image-generation-based methods often constrain viewpoint control and introduce multi-view inconsistencies. In this work, we present Video Perception models for 3D Scene synthesis (VIPScene), a novel framework that exploits the encoded commonsense knowledge of the 3D physical world in video generation models to ensure coherent scene layouts and consistent object placements across views. VIPScene accepts both text and image prompts and seamlessly integrates video generation, feedforward 3D reconstruction, and open-vocabulary perception models to semantically and geometrically analyze each object in a scene. This enables flexible scene synthesis with high realism and structural consistency. For a more sufficient evaluation on coherence and plausibility, we further introduce First-Person View Score (FPVScore), utilizing a continuous first-person perspective to capitalize on the reasoning ability of multimodal large language models. Extensive experiments show that VIPScene significantly outperforms existing methods and generalizes well across diverse scenarios.

Determining the precise geographic location of an image at a global scale remains an unsolved challenge. Standard image retrieval techniques are inefficient due to the sheer volume of images (>100M) and fail when coverage is insufficient. Scalable solutions, however, involve a trade-off: global classification typically yields coarse results (10+ kilometers), while cross-view retrieval between ground and aerial imagery suffers from a domain gap and has been primarily studied on smaller regions. This paper introduces a hybrid approach that achieves fine-grained geo-localization across a large geographic expanse the size of a continent. We leverage a proxy classification task during training to learn rich feature representations that implicitly encode precise location information. We combine these learned prototypes with embeddings of aerial imagery to increase robustness to the sparsity of ground-level data. This enables direct, fine-grained retrieval over areas spanning multiple countries. Our extensive evaluation demonstrates that our approach can localize within 200m more than 68\% of queries of a dataset covering a large part of Europe. The code is publicly available at [scaling-geoloc.github.io](https://scaling-geoloc.github.io).

Physical environments and circumstances are fundamentally dynamic, yet current 3D datasets and evaluation benchmarks tend to concentrate on either dynamic scenarios or dynamic situations in isolation, resulting in incomplete comprehension. To overcome these constraints, we introduce Situat3DChange, an extensive dataset supporting three situation-aware change understanding tasks following the perception-action model: 121K question-answer pairs, 36K change descriptions for perception tasks, and 17K rearrangement instructions for the action task. To construct this large-scale dataset, Situat3DChange leverages 11K human observations of environmental changes to establish shared mental models and shared situational awareness for human-AI collaboration. These observations, enriched with egocentric and allocentric perspectives as well as categorical and coordinate spatial relations, are integrated using an LLM to support understanding of situated changes. To address the challenge of comparing pairs of point clouds from the same scene with minor changes, we propose SCReasoner, an efficient 3D MLLM approach that enables effective point cloud comparison with minimal parameter overhead and no additional tokens required for the language decoder. Comprehensive evaluation on Situat3DChange tasks highlights both the progress and limitations of MLLMs in dynamic scene and situation understanding. Additional experiments on data scaling and cross-domain transfer demonstrate the task-agnostic effectiveness of using Situat3DChange as a training dataset for MLLMs. The established dataset and source code are publicly available at: https://github.com/RuipingL/Situat3DChange.

Learning effective multi-modal 3D representations of objects is essential for numerous applications, such as augmented reality and robotics. Existing methods often rely on task-specific embeddings that are tailored either for semantic understanding or geometric reconstruction. As a result, these embeddings typically cannot be decoded into explicit geometry and simultaneously reused across tasks. In this paper, we propose Object-X, a versatile multi-modal object representation framework capable of encoding rich object embeddings (e.g., images, point cloud, text) and decoding them back into detailed geometric and visual reconstructions. Object-X operates by geometrically grounding the captured modalities in a 3D voxel grid and learning an unstructured embedding fusing the information from the voxels with the object attributes. The learned embedding enables 3D Gaussian Splatting-based object reconstruction, while also supporting a range of downstream tasks, including scene alignment, single-image 3D object reconstruction, and localization. Evaluations on two challenging real-world datasets demonstrate that Object-X produces high-fidelity novel-view synthesis comparable to standard 3D Gaussian Splatting, while significantly improving geometric accuracy. Moreover, Object-X achieves competitive performance with specialized methods in scene alignment and localization. Critically, our object-centric descriptors require 3-4 orders of magnitude less storage compared to traditional image- or point cloud-based approaches, establishing Object-X as a scalable and highly practical solution for multi-modal 3D scene representation.

We propose an online 3D semantic segmentation method that incrementally reconstructs a 3D semantic map from a stream of RGB-D frames. Unlike offline methods, ours is directly applicable to scenarios with real-time constraints, such as robotics or mixed reality. To overcome the inherent challenges of online methods, we make two main contributions. First, to effectively extract information from the input RGB-D video stream, we jointly estimate geometry and semantic labels per frame in 3D. A key focus of our approach is to reason about semantic entities both in the 2D input and the local 3D domain to leverage differences in spatial context and network architectures. Our method predicts 2D features using an off-the-shelf segmentation network. The extracted 2D features are refined by a lightweight 3D network to enable reasoning about the local 3D structure. Second, to efficiently deal with an infinite stream of input RGB-D frames, a subsequent network serves as a temporal expert predicting the incremental scene updates by leveraging 2D, 3D, and past information in a learned manner. These updates are then integrated into a global scene representation. Using these main contributions, our method can enable scenarios with real-time constraints and can scale to arbitrary scene sizes by processing and updating the scene only in a local region defined by the new measurement. Our experiments demonstrate improved results compared to existing online methods that purely operate in local regions and show that complementary sources of information can boost the performance. We provide a thorough ablation study on the benefits of different architectural as well as algorithmic design decisions. Our method yields competitive results on the popular ScanNet benchmark and SceneNN dataset.

State-of-the-art research of traditional computer vision is increasingly leveraged in the surgical domain. A particular focus in computer-assisted surgery is to replace marker-based tracking systems for instrument localization with pure image-based 6DoF pose estimation using deep-learning methods. However, state-of-the-art single-view pose estimation methods do not yet meet the accuracy required for surgical navigation. In this context, we investigate the benefits of multi-view setups for highly accurate and occlusion-robust 6DoF pose estimation of surgical instruments and derive recommendations for an ideal camera system that addresses the challenges in the operating room. Our contributions are threefold. First, we present a multi-view RGB-D video dataset of ex-vivo spine surgeries, captured with static and head-mounted cameras and including rich annotations for surgeon, instruments, and patient anatomy. Second, we perform an extensive evaluation of three state-of-the-art single-view and multi-view pose estimation methods, analyzing the impact of camera quantities and positioning, limited real-world data, and static, hybrid, or fully mobile camera setups on the pose accuracy, occlusion robustness, and generalizability. Third, we design a multi-camera system for marker-less surgical instrument tracking, achieving an average position error of 1.01mm and orientation error of 0.89° for a surgical drill, and 2.79mm and 3.33° for a screwdriver under optimal conditions. Our results demonstrate that marker-less tracking of surgical instruments is becoming a feasible alternative to existing marker-based systems.

Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this article, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale.

To reconstruct the 3D geometry from calibrated images, learning-based multi-view stereo (MVS) methods typically perform multi-view depth estimation and then fuse depth maps into a mesh or point cloud. To improve the computational efficiency, many methods initialize a coarse depth map and then gradually refine it in higher resolutions. Recently, diffusion models achieve great success in generation tasks. Starting from a random noise, diffusion models gradually recover the sample with an iterative denoising process. In this paper, we propose a novel MVS framework, which introduces diffusion models in MVS. Specifically, we formulate depth refinement as a conditional diffusion process. Considering the discriminative characteristic of depth estimation, we design a condition encoder to guide the diffusion process. To improve efficiency, we propose a novel diffusion network combining lightweight 2D U-Net and convolutional GRU. Moreover, we propose a novel confidence-based sampling strategy to adaptively sample depth hypotheses based on the confidence estimated by diffusion model. Based on our novel MVS framework, we propose two novel MVS methods, DiffMVS and CasDiffMVS. DiffMVS achieves competitive performance with state-of-the-art efficiency in run-time and GPU memory. CasDiffMVS achieves state-of-the-art performance on DTU, Tanks & Temples and ETH3D.

We propose an approach for estimating the relative pose between calibrated image pairs by jointly exploiting points, lines, and their coincidences in a hybrid manner. We investigate all possible configurations where these data modalities can be used together and review the minimal solvers available in the literature. Our hybrid framework combines the advantages of all configurations, enabling robust and accurate estimation in challenging environments. In addition, we design a method for jointly estimating multiple vanishing point correspondences in two images, and a bundle adjustment that considers all relevant data modalities. Experiments on various indoor and outdoor datasets show that our approach outperforms point-based methods, improving AUC@ 10° by 1-7 points while running at comparable speeds. The source code of the solvers and hybrid framework will be made public.

Point cloud registration has seen recent success with several learning-based methods that focus on correspondence matching, and as such, optimize only for this objective. Following the learning step of correspondence matching, they evaluate the estimated rigid transformation with a RANSAC-like framework. While it is an indispensable component of these methods, it prevents a fully end-to-end training, leaving the objective to minimize the pose error non-served. We present a novel solution, Q-REG, which utilizes rich geometric information to estimate the rigid pose from a single correspondence. Q-REG allows to formalize the robust estimation as an exhaustive search, hence enabling end-to-end training that optimizes over both objectives of correspondence matching and rigid pose estimation. We demonstrate in the experiments that Q-REG is agnostic to the correspondence matching method and provides consistent improvement both when used only in inference and in end-to-end training. It sets a new state-of-the-art on the 3DMatch, KITTI and ModelNet benchmarks. Our code and models are available at https://github.com/jinsz/Q-REG.

Semantic annotations are indispensable to train or evaluate perception models, yet very costly to acquire. This work introduces a fully automated 2D/3D labeling framework that, without any human intervention, can generate labels for RGB-D scans at equal (or better) level of accuracy than comparable manually annotated datasets such as ScanNet. Our approach is based on an ensemble of state-of-the-art segmentation models and 3D lifting through neural rendering. We demonstrate the effectiveness of our LabelMaker pipeline by generating significantly better labels for the ScanNet datasets and automatically labelling the previously unlabeled ARKitScenes dataset. Code and models are available at labelmaker.org labelmaker.org.

Neural implicit representations have recently become popular in simultaneous localization and mapping (SLAM), especially in dense visual SLAM. However, existing works either rely on RGB-D sensors or require a separate monocular SLAM approach for camera tracking, and fail to produce high-fidelity 3D dense reconstructions. To address these shortcomings, we present NICER-SLAM, a dense RGB SLAM system that simultaneously optimizes for camera poses and a hierarchical neural implicit map representation, which also allows for high-quality novel view synthesis. To facilitate the optimization process for mapping, we integrate additional supervision signals including easy-to-obtain monocular geometric cues and optical flow, and also introduce a simple warping loss to further enforce geometric consistency. Moreover, to further boost performance in complex large-scale scenes, we also propose a local adaptive transformation from signed distance functions (SDFs) to density in the volume rendering equation. On multiple challenging indoor and outdoor datasets, NICER-SLAM demonstrates strong performance in dense mapping, novel view synthesis, and tracking, even competitive with recent RGB-D SLAM systems. Project page: https://nicer-slam.github.io/.

We estimate the radiance field of large-scale dynamic ar-eas from multiple vehicle captures under varying environ-mental conditions. Previous works in this domain are ei-ther restricted to static environments, do not scale to more than a single short video, or struggle to separately repre-sent dynamic object instances. To this end, we present a novel, decomposable radiance field approach for dynamic urban environments. We propose a multi-level neural scene graph representation that scales to thousands of images from dozens of sequences with hundreds of fast-moving ob-jects. To enable efficient training and rendering of our rep-resentation, we develop a fast composite ray sampling and rendering scheme. To test our approach in urban driving scenarios, we introduce a new, novel view synthesis bench-mark. We show that our approach outperforms prior art by a significant margin on both established and our proposed benchmark while being faster in training and rendering.

Recently, we have witnessed the explosive growth of various volumetric representations in modeling animatable head avatars. However, due to the diversity of frameworks, there is no practical method to support high-level applications like 3D head avatar editing across different representations. In this paper, we propose a generic avatar editing approach that can be universally applied to various 3DMM-driving volumetric head avatars. To achieve this goal, we design a novel expression-aware modification generative model, which enables lift 2D editing from a single image to a consistent 3D modification field. To ensure the effectiveness of the generative modification process, we develop several techniques, including an expression-dependent modification distillation scheme to draw knowledge from the large-scale head avatar model and 2D facial texture editing tools, implicit latent space guidance to enhance model convergence, and a segmentation-based loss reweight strategy for fine-grained texture inversion. Extensive experiments demonstrate that our method delivers high-quality and consistent results across multiple expression and viewpoints. Project page: https://zju3dv.github.io/geneavatar/.

Visual odometry estimates the motion of a moving cam-era based on visual input. Existing methods, mostly focusing on two-view point tracking, often ignore the rich tempo-ral context in the image sequence, thereby overlooking the global motion patterns and providing no assessment of the full trajectory reliability. These shortcomings hinder per-formance in scenarios with occlusion, dynamic objects, and low-texture areas. To address these challenges, we present the Long-term Effective Any Point Tracking (LEAP) mod-ule. LEAP innovatively combines visual, inter-track, and temporal cues with mindfully selected anchors for dynamic track estimation. Moreover, LEAP's temporal probabilistic formulation integrates distribution updates into a learnable iterative refinement module to reason about point-wise un-certainty. Based on these traits, we develop LEAP-VO, a robust visual odometry system adept at handling occlusions and dynamic scenes. Our mindful integration showcases a novel practice by employing long-term point tracking as the front-end. Extensive experiments demonstrate that the pro-posed pipeline significantly outperforms existing baselines across various visual odometry benchmarks.

In this paper we propose an efficient data-driven solution to self-localization within a floorplan. Floorplan data is readily available, long-term persistent and inherently robust to changes in the visual appearance. Our method does not require retraining per map and location or demand a large database of images of the area of interest. We propose a novel probabilistic model consisting of an observation and a novel temporal filtering module. Operating internally with an efficient ray-based representation, the observation module consists of a single and a multiview module to predict horizontal depth from images and fuses their results to benefit from advantages offered by either methodology. Our method operates on conventional consumer hardware and overcomes a common limitation of competing methods [16], [17], [20], [28] that often demand upright images. Our full system meets real-time requirements, while outperforming the state-of-the-art [20], [28] by a significant margin.

Existing 3D scene understanding methods are heavily focused on 3D semantic and instance segmentation. However, identifying objects and their parts only constitutes an intermediate step towards a more fine-grained goal, which is effectively interacting with the functional interactive elements (e.g., handles, knobs, buttons) in the scene to accomplish diverse tasks. To this end, we introduce SceneFun3D, a large-scale dataset with more than 14. Bk highly accurate interaction annotations for 710 high-resolution real-world 3D indoor scenes. We accompany the annotations with motion parameter information, describing how to interact with these elements, and a diverse set of natural language descriptions of tasks that involve manipulating them in the scene context. To showcase the value of our dataset, we introduce three novel tasks, namely functionality segmentation, task-driven affordance grounding and 3D motion estimation, and adapt existing state-of-the-art methods to tackle them. Our experiments show that solving these tasks in real 3D scenes remains challenging despite recent progress in closed-set and open-set 3D scene understanding methods.

We revisit certain problems of pose estimation based on 3D-2D correspondences between features which may be points or lines. Specifically, we address the two previously-studied minimal problems of estimating camera extrinsics from$p\in\{1,2\}$point-point correspondences and$l=3-p$line-line correspondences. To the best of our knowledge, all of the previously-known practical solutions to these problems required computing the roots of degree$\geq 4$(univariate) polynomials when$p=2$, or degree$\geq 8$polynomials when$p=1$. We describe and implement two elementary solutions which reduce the degrees of the needed polynomials from 4 to 2 and from 8 to 4, respectively. We show experimentally that the resulting solvers are numerically stable and fast: when compared to the previous state-of-the art, we may obtain nearly an order of magnitude speedup. The code is available at https://github.com/petrhruby97/efficient_absolute

We introduce a novel framework for multiway point cloud mosaicking (named Wednesday), designed to co-align sets of partially overlapping point clouds - typically obtained from 3D scanners or moving RGB-D cameras - into a unified coordinate system. At the core of our approach is ODIN, a learned pairwise registration algorithm that iteratively identifies overlaps and refines attention scores, employing a diffusion-based process for denoising pairwise correlation matrices to enhance matching accuracy. Further steps include constructing a pose graph from all point clouds, performing rotation averaging, a novel robust algorithm for re-estimating translations optimally in terms of consensus maximization and translation optimization. Finally, the point cloud rotations and positions are optimized jointly by a diffusion-based approach. Tested on four diverse, large-scale datasets, our method achieves state-of-the-art pairwise and multiway registration results by a large margin on all benchmarks. Our code and models are available at https://github.com/jinsz/Multiway-Point-Cloud-Mosaicking-with-Diffusion-and-Global-Optimization.

Recovering the 3D scene geometry from a single view is a fundamental yet ill-posed problem in computer vision. While classical depth estimation methods infer only a 2.5D scene representation limited to the image plane, recent approaches based on radiance fields reconstruct a full 3D representation. However, these methods still struggle with occluded regions since inferring geometry without visual observation requires (i) semantic knowledge of the surroundings, and (ii) reasoning about spatial context. We propose KYN, a novel method for single-view scene reconstruction that reasons about semantic and spatial context to predict each point's density. We introduce a vision-language modulation module to enrich point features with fine-grained semantic information. We aggregate point representations across the scene through a language-guided spatial attention mechanism to yield per-point density predictions aware of the 3D semantic context. We show that KYN improves 3D shape recovery compared to predicting density for each 3D point in isolation. We achieve state-of-the-art results in scene and object reconstruction on KITTI-360, and show improved zero-shot generalization compared to prior work. Project page: https://ruili3.github.io/kyn.

Understanding the world in first-person view is fundamental in Augmented Reality (AR). This immersive perspective brings dramatic visual changes and unique challenges compared to third-person views. Synthetic data has empowered third-person-view vision models, but its application to embodied egocentric perception tasks remains largely unexplored. A critical challenge lies in simulating natural human movements and behaviors that effectively steer the embodied cameras to capture a faithful egocentric representation of the 3D world. To address this challenge, we introduce EgoGen, a new synthetic data generator that can produce accurate and rich ground-truth training data for egocentric perception tasks. At the heart of EgoGen is a novel human motion synthesis model that directly leverages egocentric visual inputs of a virtual human to sense the 3D environment. Combined with collision-avoiding motion primitives and a two-stage reinforcement learning approach, our motion synthesis model offers a closed-loop solution where the embodied perception and movement of the virtual human are seamlessly coupled. Compared to previous works, our model eliminates the need for a predefined global path, and is directly applicable to dynamic environments. Combined with our easy-to-use and scalable data generation pipeline, we demonstrate EgoGen's efficacy in three tasks: mapping and localization for head-mounted cameras, egocentric camera tracking, and human mesh recovery from egocentric views. EgoGen will be fully open-sourced, offering a practical solution for creating realistic egocentric training data and aiming to serve as a useful tool for egocentric computer vision research.

Directly generating scenes from satellite imagery offers exciting possibilities for integration into applications like games and map services. However, challenges arise from significant view changes and scene scale. Previous efforts mainly focused on image or video generation, lacking exploration into the adaptability of scene generation for ar-bitrary views. Existing 3D generation works either oper-ate at the object level or are difficult to utilize the geometry obtained from satellite imagery. To overcome these limitations, we propose a novel architecture for direct 3D scene generation by introducing diffusion models into 3D sparse representations and combining them with neural rendering techniques. Specifically, our approach generates texture colors at the point level for a given geometry using a 3D diffusion model first, which is then transformed into a scene representation in a feed-forward manner. The representation can be utilized to render arbitrary views which would excel in both single-frame quality and inter-frame consistency. Experiments in two city-scale datasets show that our model demonstrates proficiency in generating photorealistic street-view image sequences and cross-view urban scenes from satellite imagery.

Real-world objects and environments are predominantly composed of edge features, including straight lines and curves. Such edges are crucial elements for various applications, such as CAD modeling, surface meshing, lane mapping, etc. However, existing traditional methods only prioritize lines over curves for simplicity in geometric modeling. To this end, we introduce EMAP, a new method for learning 3D edge representations with a focus on both lines and curves. Our method implicitly encodes 3D edge distance and direction in Unsigned Distance Functions (UDF) from multi-view edge maps. On top of this neural representation, we propose an edge extraction algorithm that robustly abstracts parametric 3D edges from the inferred edge points and their directions. Comprehensive evaluations demonstrate that our method achieves better 3D edge reconstruction on multiple challenging datasets. We further show that our learned UDF field enhances neural surface reconstruction by capturing more details.

Neural Radiance Fields (NeRFs) have shown remarkable success in synthesizing photorealistic views from multi-view images of static scenes, but face challenges in dynamic, real-world environments with distractors like moving objects, shadows, and lighting changes. Existing methods manage controlled environments and low occlusion ra-tios but fall short in render quality, especially under high occlusion scenarios. In this paper, we introduce NeRF On-the-go, a simple yet effective approach that enables the ro-bust synthesis of novel views in complex, in-the-wild scenes from only casually captured image sequences. Delving into uncertainty, our method not only efficiently eliminates dis-tractors, even when they are predominant in captures, but also achieves a notably faster convergence speed. Through comprehensive experiments on various scenes, our method demonstrates a significant improvement over state-of-the-art techniques. This advancement opens new avenues for NeRF in diverse and dynamic real-world applications.

Scene coordinate regression (SCR) methods are a family of visual localization methods that directly regress 2D-3D matches for camera pose estimation. They are effective in small-scale scenes but face significant challenges in large-scale scenes that are further amplified in the absence of ground truth 3D point clouds for supervision. Here, the model can only rely on reprojection constraints and needs to implicitly triangulate the points. The challenges stem from a fundamental dilemma: The network has to be invariant to observations of the same landmark at different viewpoints and lighting conditions, etc., but at the same time discriminate unrelated but similar observations. The latter becomes more relevant and severe in larger scenes. In this work, we tackle this problem by introducing the concept of co-visibility to the network. We propose GLACE, which integrates pre-trained global and local encodings and enables SCR to scale to large scenes with only a single small-sized network. Specifically, we propose a novel feature diffusion technique that implicitly groups the reprojection constraints with co-visibility and avoids overfitting to trivial solutions. Additionally, our position decoder parameterizes the output positions for large-scale scenes more effectively. Without using 3D models or depth maps for supervision, our method achieves state-of-the-art results on large-scale scenes with a low-map-size model. On Cambridge landmarks, with a single model, we achieve 17% lower median position error than Poker, the ensemble variant of the state-of-the-art SCR method ACE. Code is avail-able at: https://github.com/cvg/glace.

We present Multi-Baseline Radiance Fields (MuRF), a general feed-forward approach to solving sparse view syn-thesis under multiple different baseline settings (small and large baselines, and different number of input views). To render a target novel view, we discretize the 3D space into planes parallel to the target image plane, and accordingly construct a target view frustum volume. Such a target volume representation is spatially aligned with the target view, which effectively aggregates relevant information from the input views for high-quality rendering. It also facilitates subsequent radiance field regression with a convolutional network thanks to its axis-aligned nature. The 3D context modeled by the convolutional network enables our method to synthesis sharper scene structures than prior works. Our MuRF achieves state-of-the-art performance across multiple different baseline settings and diverse scenarios ranging from simple objects (DTU) to complex indoor and outdoor scenes (RealEstatel0K and LLFF). We also show promising zero-shot generalization abilities on the Mip-NeRF 360 dataset, demonstrating the general applicability of MuRF.

We propose SNI-SLAM, a semantic SLAM system utilizing neural implicit representation, that simultaneously performs accurate semantic mapping, high-quality surface reconstruction, and robust camera tracking. In this system, we introduce hierarchical semantic representation to allow multi-level semantic comprehension for top-down structured semantic mapping of the scene. In addition, to fully utilize the correlation between multiple attributes of the environment, we integrate appearance, geometry and semantic features through cross-attention for feature collaboration. This strategy enables a more multifaceted understanding of the environment, thereby allowing SNI-SLAM to remain robust even when single attribute is defective. Then, we design an internal fusion-based decoder to obtain semantic, RGB, Truncated Signed Distance Field (TSDF) values from multi-level features for accurate decoding. Furthermore, we propose a feature loss to update the scene representation at the feature level. Compared with low-level losses such as RGB loss and depth loss, our feature loss is capable of guiding the network optimization on a higher-level. Our SNI-SLAM method demonstrates superior performance over all recent NeRF-based SLAM methods in terms of mapping and tracking accuracy on Replica and ScanNet datasets, while also showing excellent capabilities in accurate semantic segmentation and real-time semantic mapping. Codes will be available at https://github.com/IRMVLab/SNI-SLAM.

Neural fields, a category of neural networks trained to represent high-frequency signals, have gained significant attention in recent years due to their impressive performance in modeling complex 3D data, such as signed distance (SDFs) or radiance fields (NeRFs), via a single multi-layer perceptron (MLP). However, despite the power and simplicity of representing signals with an MLP, these methods still face challenges when modeling large and complex temporal signals due to the limited capacity of MLPs. In this paper, we propose an effective approach to address this limitation by incorporating temporal residual layers into neural fields, dubbed ResFields. It is a novel class of networks specifically designed to effectively represent complex temporal signals. We conduct a comprehensive analysis of the properties of ResFields and propose a matrix factorization technique to reduce the number of trainable parameters and enhance generalization capabilities. Importantly, our formulation seamlessly integrates with existing MLP-based neural fields and consistently improves results across various challenging tasks: 2D video approximation, dynamic shape modeling via temporal SDFs, and dynamic NeRF reconstruction. Lastly, we demonstrate the practical utility of ResFields by showcasing its effectiveness in capturing dynamic 3D scenes from sparse RGBD cameras of a lightweight capture system.

As a promising fashion for visual localization, scene coordinate regression (SCR) has seen tremendous progress in the past decade. Most recent methods usually adopt neural networks to learn the mapping from image pixels to 3D scene coordinates, which requires a vast amount of annotated training data. We propose to leverage Neural Radiance Fields (NeRF) to generate training samples for SCR. Despite NeRF’s efficiency in rendering, many of the rendered data are polluted by artifacts or only contain minimal information gain, which can hinder the regression accuracy or bring unnecessary computational costs with redundant data. These challenges are addressed in three folds in this paper: (1) A NeRF is designed to separately predict uncertainties for the rendered color and depth images, which reveal data reliability at the pixel level. (2) SCR is formulated as deep evidential learning with epistemic uncertainty, which is used to evaluate information gain and scene coordinate quality. (3) Based on the three arts of uncertainties, a novel view selection policy is formed that significantly improves data efficiency. Experiments on public datasets demonstrate that our method could select the samples that bring the most information gain and promote the performance with the highest efficiency.

This paper presents a mixed-reality human-robot teaming system. It allows human operators to see in real-time where robots are located, even if they are not in line of sight. The operator can also visualize the map that the robots create of their environment and can easily send robots to new goal positions. The system mainly consists of a mapping and a control module. The mapping module is a real-time multi-agent visual SLAM system that co-localizes all robots and mixed-reality devices to a common reference frame. Visualizations in the mixed-reality device then allow operators to see a virtual life-sized representation of the cumulative 3D map overlaid onto the real environment. As such, the operator can effectively "see through" walls into other rooms. To control robots and send them to new locations, we propose a drag-and-drop interface. An operator can grab any robot hologram in a 3D mini map and drag it to a new desired goal pose. We validate the proposed system through a user study and real-world deployments. We make the mixed-reality application publicly available at github.com/cvg/HoloLens_ros.

Rather than having each newly deployed robot create its own map of its surroundings, the growing availability of SLAM-enabled devices provides the option of simply localizing in a map of another robot or device. In cases such as multi-robot or human-robot collaboration, localizing all agents in the same map is even necessary. However, localizing e.g. a ground robot in the map of a drone or head-mounted MR headset presents unique challenges due to viewpoint changes. This work investigates how active visual localization can be used to overcome such challenges of viewpoint changes. Specifically, we focus on the problem of selecting the optimal viewpoint at a given location. We compare existing approaches in the literature with additional proposed baselines and propose a novel data-driven approach. The result demonstrates the superior performance of our data-driven approach when compared to existing methods, both in controlled simulation experiments and real-world deployment.

We introduce a new algorithm that utilizes semantic information to enhance feature matching in visual SLAM pipelines. The proposed method constructs a high-dimensional semantic descriptor for each detected ORB feature. When integrated with traditional visual ones, these descriptors aid in establishing accurate tentative point correspondences between consecutive frames. Additionally, our semantic descriptors enrich 3D map points, enhancing loop closure detection by providing deeper insights into the underlying map regions. Experiments on public large-scale datasets demonstrate that our technique surpasses the accuracy of established methods. Importantly, given its detector-agnostic nature, our algorithm also amplifies the efficacy of modern keypoint detectors, such as SuperPoint. The implementation of our algorithm can be found on Github3.

To enable self-driving vehicles accurate detection and tracking of surrounding objects is essential. While Light Detection and Ranging (LiDAR) sensors have set the benchmark for high-performance systems, the appeal of camera-only solutions lies in their cost-effectiveness. Notably, despite the prevalent use of Radio Detection and Ranging (RADAR) sensors in automotive systems, their potential in 3D detection and tracking has been largely disregarded due to data sparsity and measurement noise. As a recent development, the combination of RADARs and cameras is emerging as a promising solution. This paper presents Camera-RADAR 3D Detection and Tracking (CR3DT), a camera-RADAR fusion model for 3D object detection, and Multi-Object Tracking (MOT). Building upon the foundations of the State-of-the-Art (SotA) camera-only BEVDet architecture, CR3DT demonstrates substantial improvements in both detection and tracking capabilities, by incorporating the spatial and velocity information of the RADAR sensor. Experimental results demonstrate an absolute improvement in detection performance of 5.3% in mean Average Precision (mAP) and a 14.9% increase in Average Multi-Object Tracking Accuracy (AMOTA) on the nuScenes dataset when leveraging both modalities. CR3DT bridges the gap between high-performance and cost-effective perception systems in autonomous driving, by capitalizing on the ubiquitous presence of RADAR in automotive applications. The code is available at: https://github.com/ETH-PBL/CR3DT.

Reconstructing compositional 3D representations of scenes, where each object is represented with its own 3D model, is a highly desirable capability in robotics and augmented reality. However, most existing methods rely heavily on strong appearance priors for object discovery, therefore only working on those classes of objects on which the method has been trained, or do not allow for object manipulation, which is necessary to scan objects fully and to guide object discovery in challenging scenarios. We address these limitations with a novel interaction-guided and class-agnostic method based on object displacements that allows a user to move around a scene with an RGB-D camera, hold up objects, and finally outputs one 3D model per held-up object. Our main contribution to this end is a novel approach to detecting user-object interactions and extracting the masks of manipulated objects. On a custom-captured dataset, our pipeline discovers manipulated objects with 78.3% precision at 100% recall and reconstructs them with a mean chamfer distance of 0.90cm. Compared to Co-Fusion, the only comparable interaction-based and class-agnostic baseline, this corresponds to a reduction in chamfer distance of 73% while detecting 99% fewer false positives.

We introduce an online 2D-to-3D semantic instance mapping algorithm aimed at generating comprehensive, accurate, and efficient semantic 3D maps suitable for autonomous agents in unstructured environments. The proposed approach is based on a Voxel-TSDF representation used in recent algorithms. It introduces novel ways of integrating semantic prediction confidence during mapping, producing semantic and instance-consistent 3D regions. Further improvements are achieved by graph optimization-based semantic labeling and instance refinement. The proposed method achieves accuracy superior to the state of the art on public large-scale datasets, improving on a number of widely used metrics. We also highlight a downfall in the evaluation of recent studies: using the ground truth trajectory as input instead of a SLAM-estimated one substantially affects the accuracy, creating a large gap between the reported results and the actual performance on real-world data. The code is available: https://github.com/y9miao/ConsistentPanopticSLAM.

We present an efficient neural 3D scene representation for novel-view synthesis (NVS) in large-scale, dynamic urban areas. Existing works are not well suited for applications like mixed-reality or closed-loop simulation due to their limited visual quality and non-interactive rendering speeds. Recently, rasterization-based approaches have achieved high-quality NVS at impressive speeds. However, these methods are limited to small-scale, homogeneous data, i.e. they cannot handle severe appearance and geometry variations due to weather, season, and lighting and do not scale to larger, dynamic areas with thousands of images. We propose 4DGF, a neural scene representation that scales to large-scale dynamic urban areas, handles heterogeneous input data, and substantially improves rendering speeds. We use 3D Gaussians as an efficient geometry scaffold while relying on neural fields as a compact and flexible appearance model. We integrate scene dynamics via a scene graph at global scale while modeling articulated motions on a local level via deformations. This decomposed approach enables flexible scene composition suitable for real-world applications. In experiments, we surpass the state-of-the-art by over 3 dB in PSNR and more than 200x in rendering speed.

While the field of 3D scene reconstruction is dominated by NeRFs due to their photorealistic quality, 3D Gaussian Splatting (3DGS) has recently emerged, offering similar quality with real-time rendering speeds. However, both methods primarily excel with well-controlled 3D scenes, while in-the-wild data - characterized by occlusions, dynamic objects, and varying illumination - remains challenging. NeRFs can adapt to such conditions easily through per-image embedding vectors, but 3DGS struggles due to its explicit representation and lack of shared parameters. To address this, we introduce WildGaussians, a novel approach to handle occlusions and appearance changes with 3DGS. By leveraging robust DINO features and integrating an appearance modeling module within 3DGS, our method achieves state-of-the-art results. We demonstrate that WildGaussians matches the real-time rendering speed of 3DGS while surpassing both 3DGS and NeRF baselines in handling in-the-wild data, all within a simple architectural framework.

Neural 3D scene representations have shown great potential for 3D reconstruction from 2D images. However, reconstructing real-world captures of complex scenes still remains a challenge. Existing generic 3D reconstruction methods often struggle to represent fine geometric details and do not adequately model reflective surfaces of large-scale scenes. Techniques that explicitly focus on reflective surfaces can model complex and detailed reflections by exploiting better reflection parameterizations. However, we observe that these methods are often not robust in real scenarios where non-reflective as well as reflective components are present. In this work, we propose UniSDF, a general purpose 3D reconstruction method that can reconstruct large complex scenes with reflections. We investigate both camera view as well as reflected view-based color parameterization techniques and find that explicitly blending these representations in 3D space enables reconstruction of surfaces that are more geometrically accurate, especially for reflective surfaces. We further combine this representation with a multi-resolution grid backbone that is trained in a coarse-to-fine manner, enabling faster reconstructions than prior methods. Extensive experiments on object-level datasets DTU, Shiny Blender as well as unbounded datasets Mip-NeRF 360 and Ref-NeRF real demonstrate that our method is able to robustly reconstruct complex large-scale scenes with fine details and reflective surfaces, leading to the best overall performance. Project page: https://fangjinhuawang.github.io/UniSDF.

LiDAR point cloud semantic segmentation enables the robots to obtain fine-grained semantic information of the surrounding environment. Recently, many works project the point cloud onto the 2D image and adopt the 2D Convolutional Neural Networks (CNNs) or vision transformer for LiDAR point cloud semantic segmentation. However, since more than one point can be projected onto the same 2D position but only one point can be preserved, the previous 2D projection-based segmentation methods suffer from inevitable quantized information loss, which results in incomplete geometric structure, especially for small objects. To avoid quantized information loss, in this paper, we propose a novel spherical frustum structure, which preserves all points projected onto the same 2D position. Additionally, a hash-based representation is proposed for memory-efficient spherical frustum storage. Based on the spherical frustum structure, the Spherical Frustum sparse Convolution (SFC) and Frustum Farthest Point Sampling (F2PS) are proposed to convolve and sample the points stored in spherical frustums respectively. Finally, we present the Spherical Frustum sparse Convolution Network (SFCNet) to adopt 2D CNNs for LiDAR point cloud semantic segmentation without quantized information loss. Extensive experiments on the SemanticKITTI and nuScenes datasets demonstrate that our SFCNet outperforms previous 2D projection-based semantic segmentation methods based on conventional spherical projection and shows better performance on small object segmentation by preserving complete geometric structure. Codes will be available at https://github.com/IRMVLab/SFCNet.

Established surgical navigation systems for pedicle screw placement have been proven to be accurate, but still reveal limitations in registration or surgical guidance. Registration of preoperative data to the intraoperative anatomy remains a time-consuming, error-prone task that includes exposure to harmful radiation. Surgical guidance through conventional displays has well-known drawbacks, as information cannot be presented in-situ and from the surgeon’s perspective. Consequently, radiation-free and more automatic registration methods with subsequent surgeon-centric navigation feedback are desirable. In this work, we present a marker-less approach that automatically solves the registration problem for lumbar spinal fusion surgery in a radiation-free manner. A deep neural network was trained to segment the lumbar spine and simultaneously predict its orientation, yielding an initial pose for preoperative models, which then is refined for each vertebra individually and updated in real-time with GPU acceleration while handling surgeon occlusions. An intuitive surgical guidance is provided thanks to the integration into an augmented reality based navigation system. The registration method was verified on a public dataset with a median of 100% successful registrations, a median target registration error of 2.7 mm, a median screw trajectory error of 1.6°and a median screw entry point error of 2.3 mm. Additionally, the whole pipeline was validated in an ex-vivo surgery, yielding a 100% screw accuracy and a median target registration error of 1.0 mm. Our results meet clinical demands and emphasize the potential of RGB-D data for fully automatic registration approaches in combination with augmented reality guidance.

We present novel solutions to previously unsolved prob-lems of relative pose estimation from images whose calibration parameters, namely focal lengths and radial distortion, are unknown. Our approach enables metric reconstruction without modeling these parameters. The minimal case for reconstruction requires 13 points in 4 views for both the calibrated and uncalibrated cameras. We describe and implement the first solution to these minimal problems. In the calibrated case, this may be modeled as a polynomial sys-tem of equations with 3584 solutions. Despite the apparent intractability, the problem decomposes spectacularly. Each solution falls into a Euclidean symmetry class of size 16, and we can estimate 224 class representatives by solving a sequence of three subproblems with 28, 2, and 4 solutions. We highlight the relationship between internal constraints on the radial quadrifocal tensor and the relations among the principal minors of a$4\times 4$matrix. We also address the case of 4 upright cameras, where 7 points are minimal. Finally, we evaluate our approach on simulated and real data and benchmark against previous calibration-free solutions, and show that our method provides an efficient startup for an SfM pipeline with radial cameras.

In contrast to sparse keypoints, a handful of line segments can concisely encode the high-level scene layout, as they often delineate the main structural elements. In addition to offering strong geometric cues, they are also omnipresent in urban landscapes and indoor scenes. Despite their apparent advantages, current line-based reconstruction methods are far behind their point-based counterparts. In this paper we aim to close the gap by introducing LIMAP, a library for 3D line mapping that robustly and efficiently creates 3D line maps from multi-view imagery. This is achieved through revisiting the degeneracy problem of line triangulation, carefully crafted scoring and track building, and exploiting structural priors such as line coincidence, parallelism, and orthogonality. Our code integrates seamlessly with existing point-based Structure-from-Motion methods and can leverage their 3D points to further improve the line reconstruction. Furthermore, as a byproduct, the method is able to recover 3D association graphs between lines and points / vanishing points (VPs). In thorough experiments, we show that LIMAP significantly outperforms existing approaches for 3D line mapping. Our robust 3D line maps also open up new research directions. We show two example applications: visual localization and bundle adjustment, where integrating lines alongside points yields the best results. Code is available at https://github.com/cvg/limap.

Line segments are ubiquitous in our human-made world and are increasingly used in vision tasks. They are complementary to feature points thanks to their spatial extent and the structural information they provide. Traditional line detectors based on the image gradient are extremely fast and accurate, but lack robustness in noisy images and challenging conditions. Their learned counterparts are more repeatable and can handle challenging images, but at the cost of a lower accuracy and a bias towards wireframe lines. We propose to combine traditional and learned approaches to get the best of both worlds: an accurate and robust line detector that can be trained in the wild without ground truth lines. Our new line segment detector, DeepLSD, processes images with a deep network to generate a line attraction field, before converting it to a surrogate image gradient magnitude and angle, which is then fed to any existing handcrafted line detector. Additionally, we propose a new optimization tool to refine line segments based on the attraction field and vanishing points. This refinement improves the accuracy of current deep detectors by a large margin. We demonstrate the performance of our method on low-level line detection metrics, as well as on several downstream tasks using multiple challenging datasets. The source code and models are available at https://github.com/cvg/DeepLSD.

Traditional 3D scene understanding approaches rely on labeled 3D datasets to train a model for a single task with supervision. We propose OpenScene, an alternative approach where a model predicts dense features for 3D scene points that are co-embedded with text and image pixels in CLIP feature space. This zero-shot approach enables task-agnostic training and open-vocabulary queries. For example, to perform SOTA zero-shot 3D semantic segmentation it first infers CLIP features for every 3D point and later classifies them based on similarities to embeddings of arbitrary class labels. More interestingly, it enables a suite of open-vocabulary scene understanding applications that have never been done before. For example, it allows a user to enter an arbitrary text query and then see a heat map indicating which parts of a scene match. Our approach is effective at identifying objects, materials, affordances, activities, and room types in complex 3D scenes, all using a single model trained without any labeled 3D data.

The success of the Neural Radiance Fields (NeRF) in novel view synthesis has inspired researchers to propose neural implicit scene reconstruction. However, most existing neural implicit reconstruction methods optimize perscene parameters and therefore lack generalizability to new scenes. We introduce VolRecon, a novel generalizable implicit reconstruction method with Signed Ray Distance Function (SRDF). To reconstruct the scene with fine details and little noise, VolRecon combines projection features aggregated from multi-view features, and volume features interpolated from a coarse global feature volume. Using a ray transformer, we compute SRDF values of sampled points on a ray and then render color and depth. On DTU dataset, VolRecon outperforms SparseNeuS by about 30% in sparse view reconstruction and achieves comparable accuracy as MVSNet in full view reconstruction. Furthermore, our approach exhibits good generalization performance on the large-scale ETH3D benchmark. Code is available at https://github.com/IVRL/VolRecon/.

Neural Radiance Fields (NeRFs) are emerging as a ubiquitous scene representation that allows for novel view synthesis. Increasingly, NeRFs will be shareable with other people. Before sharing a NeRF, though, it might be desirable to remove personal information or unsightly objects. Such removal is not easily achieved with the current NeRF editing frameworks. We propose a framework to remove objects from a NeRF representation created from an RGBD sequence. Our NeRF inpainting method leverages recent work in 2D image inpainting and is guided by a userprovided mask. Our algorithm is underpinned by a confidence based view selection procedure. It chooses which of the individual 2D inpainted images to use in the creation of the NeRF, so that the resulting inpainted NeRF is 3D consistent. We show that our method for NeRF editing is effective for synthesizing plausible inpaintings in a multi-view coherent manner, outperforming competing methods. We validate our approach by proposing a new and still-challenging dataset for the task of NeRF inpainting.

We introduce LightGlue, a deep neural network that learns to match local features across images. We revisit multiple design decisions of SuperGlue, the state of the art in sparse matching, and derive simple but effective improvements. Cumulatively, they make LightGlue more efficient – in terms of both memory and computation, more accurate, and much easier to train. One key property is that LightGlue is adaptive to the difficulty of the problem: the inference is much faster on image pairs that are intuitively easy to match, for example because of a larger visual overlap or limited appearance change. This opens up exciting prospects for deploying deep matchers in latency-sensitive applications like 3D reconstruction. The code and trained models are publicly available at github.com/cvg/LightGlue.

Although point cloud registration has achieved remarkable advances in object-level and indoor scenes, large-scale registration methods are rarely explored. Challenges mainly arise from the huge point number, complex distribution, and outliers of outdoor LiDAR scans. In addition, most existing registration works generally adopt a two-stage paradigm: They first find correspondences by extracting discriminative local features and then leverage estimators (eg. RANSAC) to filter outliers, which are highly dependent on well-designed descriptors and post-processing choices. To address these problems, we propose an end-to-end transformer network (RegFormer) for large-scale point cloud alignment without any further post-processing. Specifically, a projection-aware hierarchical transformer is proposed to capture long-range dependencies and filter outliers by extracting point features globally. Our transformer has linear complexity, which guarantees high efficiency even for large-scale scenes. Furthermore, to effectively reduce mismatches, a bijective association transformer is designed for regressing the initial transformation. Extensive experiments on KITTI and NuScenes datasets demonstrate that our RegFormer achieves competitive performance in terms of both accuracy and efficiency. Codes are available at https://github.com/IRMVLab/RegFormer.

Robust estimation is a crucial and still challenging task, which involves estimating model parameters in noisy environments. Although conventional sampling consensus-based algorithms sample several times to achieve robustness, these algorithms cannot use data features and historical information effectively. In this paper, we propose RLSAC, a novel Reinforcement Learning enhanced SAmple Consensus framework for end-to-end robust estimation. RLSAC employs a graph neural network to utilize both data and memory features to guide exploring directions for sampling the next minimum set. The feedback of downstream tasks serves as the reward for unsupervised training. Therefore, RL-SAC can avoid differentiating to learn the features and the feedback of downstream tasks for end-to-end robust estimation. In addition, RLSAC integrates a state transition module that encodes both data and memory features. Our experimental results demonstrate that RLSAC can learn from features to gradually explore a better hypothesis. Through analysis, it is apparent that RLSAC can be easily transferred to other sampling consensus-based robust estimation tasks. To the best of our knowledge, RLSAC is also the first method that uses reinforcement learning to sample consensus for end-to-end robust estimation. We release our codes at https://github.com/IRMVLab/RLSAC.

Recent methods on privacy-preserving image-based localization use a random line parameterization to protect the privacy of query images and database maps. The lifting of points to lines effectively drops one of the two geometric constraints traditionally used with point-to-point correspondences in structure-based localization. This leads to a significant loss of accuracy for the privacy-preserving methods. In this paper, we overcome this limitation by devising a coordinate permutation scheme that allows for recovering the original point positions during pose estimation. The recovered points provide the full 2D geometric constraints and enable us to close the gap between privacy-preserving and traditional methods in terms of accuracy. Another limitation of random line methods is their vulnerability to density based 3D line cloud inversion attacks. Our method not only provides better accuracy than the original random line based approach but also provides stronger privacy guarantees against these recently proposed attacks. Extensive experiments on standard benchmark datasets demonstrate these improvements consistently across both scenarios of protecting the privacy of query images as well as the database map.

We tackle the problem of estimating a Manhattan frame, i.e. three orthogonal vanishing points, and the unknown focal length of the camera, leveraging a prior vertical direction. The direction can come from an Inertial Measurement Unit that is a standard component of recent consumer devices, e.g., smartphones. We provide an exhaustive analysis of minimal line configurations and derive two new 2-line solvers, one of which does not suffer from singularities affecting existing solvers. Additionally, we design a new non-minimal method, running on an arbitrary number of lines, to boost the performance in local optimization. Combining all solvers in a hybrid robust estimator, our method achieves increased accuracy even with a rough prior. Experiments on synthetic and real-world datasets demonstrate the superior accuracy of our method compared to the state of the art, while having comparable runtimes. We further demonstrate the applicability of our solvers for relative rotation estimation. The code is available at https://github.com/cvg/VP-Estimation-with-Prior-Gravity.

Line segments are powerful features complementary to points. They offer structural cues, robust to drastic viewpoint and illumination changes, and can be present even in texture-less areas. However, describing and matching them is more challenging compared to points due to partial occlusions, lack of texture, or repetitiveness. This paper introduces a new matching paradigm, where points, lines, and their descriptors are unified into a single wireframe structure. We propose GlueStick, a deep matching Graph Neural Network (GNN) that takes two wireframes from different images and leverages the connectivity information between nodes to better glue them together. In addition to the increased efficiency brought by the joint matching, we also demonstrate a large boost of performance when leveraging the complementary nature of these two features in a single architecture. We show that our matching strategy outperforms the state-of-the-art approaches independently matching line segments and points for a wide variety of datasets and tasks. The code is available at https://github.com/cvg/GlueStick.

Most model-free visual object tracking methods formulate the tracking task as object location estimation given by a 2D segmentation or a bounding box in each video frame. We argue that this representation is limited and instead propose to guide and improve 2D tracking with an explicit object representation, namely the textured 3D shape and 6DoF pose in each video frame. Our representation tackles a complex long-term dense correspondence problem between all 3D points on the object for all video frames, including frames where some points are invisible. To achieve that, the estimation is driven by re-rendering the input video frames as well as possible through differentiable rendering, which has not been used for tracking before. The proposed optimization minimizes a novel loss function to estimate the best 3D shape, texture, and 6DoF pose. We improve the state-of-the-art in 2D segmentation tracking on three different datasets with mostly rigid objects.

Building 3D scene graphs has recently emerged as a topic in scene representation for several embodied AI applications to represent the world in a structured and rich manner. With their increased use in solving downstream tasks (e.g., navigation and room rearrangement), can we leverage and recycle them for creating 3D maps of environments, a pivotal step in agent operation? We focus on the fundamental problem of aligning pairs of 3D scene graphs whose overlap can range from zero to partial and can contain arbitrary changes. We propose SGAligner, the first method for aligning pairs of 3D scene graphs that is robust to in-the-wild scenarios (i.e., unknown overlap – if any – and changes in the environment). We get inspired by multimodality knowledge graphs and use contrastive learning to learn a joint, multi-modal embedding space. We evaluate on the 3RScan dataset and further showcase that our method can be used for estimating the transformation between pairs of 3D scenes. Since benchmarks for these tasks are missing, we create them on this dataset. The code, benchmark, and trained models are available on the project website.

Dense 3D reconstruction and ego-motion estimation are key challenges in autonomous driving and robotics. Compared to the complex, multi-modal systems deployed today, multi-camera systems provide a simpler, low-cost alternative. However, camera-based 3D reconstruction of complex dynamic scenes has proven extremely difficult, as existing solutions often produce incomplete or incoherent results. We propose R3D3, a multi-camera system for dense 3D reconstruction and ego-motion estimation. Our approach iterates between geometric estimation that exploits spatial-temporal information from multiple cameras, and monocular depth refinement. We integrate multi-camera feature correlation and dense bundle adjustment operators that yield robust geometric depth and pose estimates. To improve reconstruction where geometric depth is unreliable, e.g. for moving objects or low-textured regions, we introduce learnable scene priors via a depth refinement network. We show that this design enables a dense, consistent 3D reconstruction of challenging, dynamic outdoor environments. Consequently, we achieve state-of-the-art dense depth prediction on the DDAD and NuScenes benchmarks.

We propose a new method, called curvature similarity extractor (CSE), for improving local feature matching across images. CSE calculates the curvature of the local 3D surface patch for each detected feature point in a viewpoint-invariant manner via fitting quadrics to predicted monocular depth maps. This curvature is then leveraged as an additional signal in feature matching with off-the-shelf matchers like SuperGlue and LoFTR. Additionally, CSE enables end-to-end joint training by connecting the matcher and depth predictor networks. Our experiments demonstrate on large-scale real-world datasets that CSE consistently improves the accuracy of state-of-the-art methods. Fine-tuning the depth prediction network further enhances the accuracy. The proposed approach achieves state-of-the-art results on the ScanNet dataset, showcasing the effectiveness of incorporating 3D geometric information into feature matching.1

Building an interactive AI assistant that can perceive, reason, and collaborate with humans in the real world has been a long-standing pursuit in the AI community. This work is part of a broader research effort to develop intelligent agents that can interactively guide humans through performing tasks in the physical world. As a first step in this direction, we introduce HoloAssist, a large-scale egocentric human interaction dataset, where two people collaboratively complete physical manipulation tasks. The task performer executes the task while wearing a mixed-reality headset that captures seven synchronized data streams. The task instructor watches the performer’s egocentric video in real time and guides them verbally. By augmenting the data with action and conversational annotations and observing the rich behaviors of various participants, we present key insights into how human assistants correct mistakes, intervene in the task completion procedure, and ground their instructions to the environment. HoloAssist spans 166 hours of data captured by 350 unique instructor-performer pairs. Furthermore, we construct and present benchmarks on mistake detection, intervention type prediction, and hand forecasting, along with detailed analysis. We expect HoloAssist will provide an important resource for building AI assistants that can fluidly collaborate with humans in the real world. Data can be downloaded at https://holoassist.github.io/.

Existing inverse rendering combined with neural rendering methods can only perform editable novel view synthesis on object-specific scenes, while we present intrinsic neural radiance fields, dubbed IntrinsicNeRF, which introduce intrinsic decomposition into the NeRF-based neural rendering method and can extend its application to room-scale scenes. Since intrinsic decomposition is a fundamentally under-constrained inverse problem, we propose a novel distance-aware point sampling and adaptive reflectance iterative clustering optimization method, which enables IntrinsicNeRF with traditional intrinsic decomposition constraints to be trained in an unsupervised manner, resulting in multi-view consistent intrinsic decomposition results. To cope with the problem that different adjacent instances of similar reflectance in a scene are incorrectly clustered together, we further propose a hierarchical clustering method with coarse-to-fine optimization to obtain a fast hierarchical indexing representation. It supports compelling real-time augmented applications such as recoloring and illumination variation. Extensive experiments and editing samples on both object-specific/room-scale scenes and synthetic/real-word data demonstrate that we can obtain consistent intrinsic decomposition results and high-fidelity novel view synthesis even for challenging sequences.

We propose a method to estimate 3D human poses from substantially blurred images. The key idea is to tackle the inverse problem of image deblurring by modeling the forward problem with a 3D human model, a texture map, and a sequence of poses to describe human motion. The blurring process is then modeled by a temporal image aggregation step. Using a differentiable renderer, we can solve the inverse problem by backpropagating the pixel-wise re-projection error to recover the best human motion representation that explains a single or multiple input images. Since the image reconstruction loss alone is insufficient, we present additional regularization terms. To the best of our knowledge, we present the first method to tackle this problem. Our method consistently outperforms other methods on significantly blurry inputs since they lack one or multiple key functionalities that our method unifies, i.e. image deblurring with sub-frame accuracy and explicit 3D modeling of non-rigid human motion.

A distinctive representation of image patches in form of features is a key component of many computer vision and robotics tasks, such as image matching, image retrieval, and visual localization. State-of-the-art descriptors, from hand-crafted descriptors such as SIFT to learned ones such as HardNet, are usually high-dimensional; 128 dimensions or even more. The higher the dimensionality, the larger the memory consumption and computational time for approaches using such descriptors. In this paper, we investigate multi-layer perceptrons (MLPs) to extract low-dimensional but high-quality descriptors. We thoroughly analyze our method in unsuper-vised, self-supervised, and supervised settings, and evaluate the dimensionality reduction results on four representative descriptors. We consider different applications, including visual localization, patch verification, image matching and retrieval. The experiments show that our lightweight MLPs trained using supervised method achieve better dimensionality reduction than PCA. The lower-dimensional descriptors generated by our approach outperform the original higher-dimensional descriptors in downstream tasks, especially for the hand-crafted ones. The code is available at https://github.com/PRBonn/descriptor-dr.

Intelligent robots require object-level scene understanding to reason about possible tasks and interactions with the environment. Moreover, many perception tasks such as scene reconstruction, image retrieval, or place recognition can benefit from reasoning on the level of objects. While keypoint-based matching can yield strong results for finding correspondences for images with small to medium view point changes, for large view point changes, matching semantically on the object-level becomes advantageous. In this paper, we propose a learning-based approach which combines local keypoints with novel object-level features for matching object detections between RGB images. We train our object-level matching features based on appearance and inter-frame and cross-frame spatial relations between objects in an associative graph neural network. We demonstrate our approach in a large variety of views on realistically rendered synthetic images. Our approach compares favorably to previous state-of-the-art object-level matching approaches and achieves improved performance over a pure keypoint-based approach for large view-point changes.

Estimating camera motion in deformable scenes poses a complex and open research challenge. Most existing non-rigid structure from motion techniques assume to observe also static scene parts besides deforming scene parts in order to establish an anchoring reference. However, this assumption does not hold true in certain relevant application cases such as endoscopies. Deformable odometry and SLAM pipelines, which tackle the most challenging scenario of exploratory trajectories, suffer from a lack of robustness and proper quantitative evaluation methodologies. To tackle this issue with a common benchmark, we introduce the Drunkard's Dataset, a challenging collection of synthetic data targeting visual navigation and reconstruction in deformable environments. This dataset is the first large set of exploratory camera trajectories with ground truth inside 3D scenes where every surface exhibits non-rigid deformations over time. Simulations in realistic 3D buildings lets us obtain a vast amount of data and ground truth labels, including camera poses, RGB images and depth, optical flow and normal maps at high resolution and quality. We further present a novel deformable odometry method, dubbed the Drunkard’s Odometry, which decomposes optical flow estimates into rigid-body camera motion and non-rigid scene deformations. In order to validate our data, our work contains an evaluation of several baselines as well as a novel tracking error metric which does not require ground truth data. Dataset and code: https://davidrecasens.github.io/TheDrunkard'sOdometry/

Semantic 2D maps are commonly used by humans and machines for navigation purposes, whether it's walking or driving. However, these maps have limitations: they lack detail, often contain inaccuracies, and are difficult to create and maintain, especially in an automated fashion. Can we use _raw imagery_ to automatically create _better maps_ that can be easily interpreted by both humans and machines? We introduce SNAP, a deep network that learns rich 2D _neural_ maps from ground-level and overhead images. We train our model to align neural maps estimated from different inputs, supervised only with camera poses over tens of millions of StreetView images. SNAP can resolve the location of challenging image queries beyond the reach of traditional methods, outperforming the state of the art in localization by a large margin. Moreover, our neural maps encode not only geometry and appearance but also high-level semantics, discovered without explicit supervision. This enables effective pre-training for data-efficient semantic scene understanding, with the potential to unlock cost-efficient creation of more detailed maps.

We introduce the task of open-vocabulary 3D instance segmentation. Current approaches for 3D instance segmentation can typically only recognize object categories from a pre-defined closed set of classes that are annotated in the training datasets. This results in important limitations for real-world applications where one might need to perform tasks guided by novel, open-vocabulary queries related to a wide variety of objects. Recently, open-vocabulary 3D scene understanding methods have emerged to address this problem by learning queryable features for each point in the scene. While such a representation can be directly employed to perform semantic segmentation, existing methods cannot separate multiple object instances. In this work, we address this limitation, and propose OpenMask3D, which is a zero-shot approach for open-vocabulary 3D instance segmentation. Guided by predicted class-agnostic 3D instance masks, our model aggregates per-mask features via multi-view fusion of CLIP-based image embeddings. Experiments and ablation studies on ScanNet200 and Replica show that OpenMask3D outperforms other open-vocabulary methods, especially on the long-tail distribution. Qualitative experiments further showcase OpenMask3D’s ability to segment object properties based on free-form queries describing geometry, affordances, and materials.

3D hand pose estimation is a challenging problem in computer vision due to the high degrees-of-freedom of hand articulated motion space and large viewpoint variation. As a consequence, similar poses observed from multiple views can be dramatically different. In order to deal with this issue, view-independent features are required to achieve state-of-the-art performance. In this paper, we investigate the impact of view-independent features on 3D hand pose estimation from a single depth image, and propose a novel recurrent neural network for 3D hand pose estimation, in which a cascaded 3D pose-guided alignment strategy is designed for view-independent feature extraction and a recurrent hand pose module is designed for modeling the dependencies among sequential aligned features for 3D hand pose estimation. In particular, our cascaded pose-guided 3D alignments are performed in 3D space in a coarse-to-fine fashion. First, hand joints are predicted and globally transformed into a canonical reference frame; Second, the palm of the hand is detected and aligned; Third, local transformations are applied to the fingers to refine the final predictions. The proposed recurrent hand pose module for aligned 3D representation can extract recurrent pose-aware features and iteratively refines the estimated hand pose. Our recurrent module could be utilized for both single-view estimation and sequence-based estimation with 3D hand pose tracking. Experiments show that our method improves the state-of-the-art by a large margin on popular benchmarks with the simple yet efficient alignment and network architectures.

Undesirable reflections contained in photos taken in front of glass windows or doors often degrade visual quality of the image. Separating two layers apart benefits both human and machine perception. The polarization status of the light changes after refraction or reflection, providing more observations of the scene, which can benefit the reflection separation. Different from previous works that take three or more polarization images as input, we propose to exploit physical constraints from a pair of unpolarized and polarized images to separate reflection and transmission layers in this paper. Due to the simplified capturing setup, the system is more under-determined compared to the existing polarization-based works. In order to solve this problem, we propose to estimate the semi-reflector orientation first to make the physical image formation well-posed, and then learn to reliably separate two layers using additional networks based on both physical and numerical analysis. In addition, a motion estimation network is introduced to handle the misalignment of paired input. Quantitative and qualitative experimental results show our approach performs favorably over existing polarization and single image based solutions.

In this paper, we propose some efficient multi-view stereo methods for accurate and complete depth map estimation. We first present our basic methods with Adaptive Checkerboard sampling and Multi-Hypothesis joint view selection (ACMH & ACMH+). Based on our basic models, we develop two frameworks to deal with the depth estimation of ambiguous regions (especially low-textured areas) from two different perspectives: multi-scale information fusion and planar geometric clue assistance. For the former one, we propose a multi-scale geometric consistency guidance framework (ACMM) to obtain the reliable depth estimates for low-textured areas at coarser scales and guarantee that they can be propagated to finer scales. For the latter one, we propose a planar prior assisted framework (ACMP). We utilize a probabilistic graphical model to contribute a novel multi-view aggregated matching cost. At last, by taking advantage of the above frameworks, we further design a multi-scale geometric consistency guided and planar prior assisted multi-view stereo (ACMMP). This greatly enhances the discrimination of ambiguous regions and helps their depth sensing. Experiments on extensive datasets show our methods achieve state-of-the-art performance, recovering the depth estimation not only in low-textured areas but also in details. Related codes are available at https://github.com/GhiXu.

Visual (re)localization addresses the problem of estimating the 6-DoF (Degree of Freedom) camera pose of a query image captured in a known scene, which is a key building block of many computer vision and robotics applications. Recent advances in structure-based localization solve this problem by memorizing the mapping from image pixels to scene coordinates with neural networks to build 2D-3D correspondences for camera pose optimization. However, such memorization requires training by amounts of posed images in each scene, which is heavy and inefficient. On the contrary, few-shot images are usually sufficient to cover the main regions of a scene for a human operator to perform visual localization. In this paper, we propose a scene region classification approach to achieve fast and effective scene memorization with few-shot images. Our insight is leveraging a) pre-learned feature extractor, b) scene region classifier, and c) meta-learning strategy to accelerate training while mitigating overfitting. We evaluate our method on both indoor and outdoor benchmarks. The experiments validate the effectiveness of our method in the few-shot setting, and the training time is significantly reduced to only a few minutes.11Code available at: https://github.com/siyandong/SRC

We propose the Model Quality Network, MQ-Net in short, for predicting the quality, e.g. the pose error of essential matrices, of models generated inside RANSAC. It replaces the traditionally used scoring techniques, e.g., inlier counting of RANSAC, truncated loss of MSAC, and the marginalization-based loss of MAGSAC++. Moreover, Minimal samples Filtering Network (MF-Net) is proposed for the early rejection of minimal samples that likely lead to degenerate models or to ones that are inconsistent with the scene geometry, e.g., due to the chirality constraint. We show on 54450 image pairs from public real-world datasets that the proposed MQ-Net leads to results superior to the state-of-the-art in terms of accuracy by a large margin. The proposed MF-Net accelerates the fundamental matrix estimation by five times and significantly reduces the essential matrix estimation time while slightly improving accuracy as well. Also, we show experimentally that consensus maximization, i.e. inlier counting, is not an inherently good measure of the model quality for relative pose estimation. The code is at https://github.com/danini/learning-goad-models-in-ransac.

Recently proposed privacy preserving solutions for cloud-based localization rely on lifting traditional point-based maps to randomized 3D line clouds. While the lifted representation is effective in concealing private information, there are two fundamental limitations. First, without careful construction of the line clouds, the representation is vulnerable to density-based inversion attacks. Secondly, after successful localization, the precise camera orientation and position is revealed to the server. However, in many scenarios, the pose itself might be sensitive information. We propose a principled approach overcoming these limitations, based on two observations. First, a full 6 DoF pose is not always necessary, and in combination with egomotion tracking even a one dimensional localization can reduce uncertainty and correct drift. Secondly, by lifting to parallel planes instead of lines, the map only provides partial constraints on the query pose, preventing the server from knowing the exact query location. If the client requires a full 6 DoF pose, it can be obtained by fusing the result from multiple queries, which can be temporally and spatially disjoint. We demonstrate the practical feasibility of this approach and show a small performance drop compared to both the conventional and privacy preserving approaches.

Temporal alignment of fine-grained human actions in videos is important for numerous applications in computer vision, robotics, and mixed reality. State-of-the-art methods directly learn image-based embedding space by leveraging powerful deep convolutional neural networks. While being straightforward, their results are far from satisfactory, the aligned videos exhibit severe temporal discontinuity without additional post-processing steps. The recent advancements in human body and hand pose estimation in the wild promise new ways of addressing the task of human action alignment in videos. In this work, based on off-the-shelf human pose estimators, we propose a novel context-aware self-supervised learning architecture to align sequences of actions. We name it CASA. Specifically, CASA employs self-attention and cross-attention mechanisms to incorporate the spatial and temporal context of human actions, which can solve the temporal dis-continuity problem. Moreover, we introduce a self-supervised learning scheme that is empowered by novel 4D augmentation techniques for 3D skeleton representations. We systematically evaluate the key components of our method. Our experiments on three public datasets demonstrate CASA significantly improves phase progress and Kendall's Tau scores over the previous state-of-the-art methods.

State-of-the-art methods for self-supervised sequential action alignment rely on deep networks that find correspondences across videos in time. They either learn frame-to-frame mapping across sequences, which does not leverage temporal information, or assume monotonic alignment between each video pair, which ignores variations in the order of actions. As such, these methods are not able to deal with common real-world scenarios that involve background frames or videos that contain non-monotonic sequence of actions. In this paper, we propose an approach to align sequential actions in the wild that involve diverse temporal variations. To this end, we propose an approach to enforce temporal priors on the optimal transport matrix, which leverages temporal consistency, while allowing for variations in the order of actions. Our model accounts for both monotonic and non-monotonic sequences and handles background frames that should not be aligned. We demonstrate that our approach consistently outperforms the state-of-the-art in self-supervised sequential action representation learning on four different benchmark datasets. Code is publicly available at https://github.com/weizheliu/VAVA.

Low-dimensional parametric models are the de-facto standard in computer vision for intrinsic camera calibration. These models explicitly describe the mapping between incoming viewing rays and image pixels. In this paper, we explore an alternative approach which implicitly models the lens distortion. The main idea is to replace the parametric model with a regularization term that ensures the latent distortion map varies smoothly throughout the image. The proposed model is effectively parameter-free and allows us to optimize the 6 degree-of-freedom camera pose without explicitly knowing the intrinsic calibration. We show that the method is applicable to a wide selection of cameras with varying distortion and in multiple applications, such as visual localization and structure-from-motion.

We propose a method for jointly estimating the 3D motion, 3D shape, and appearance of highly motion-blurred objects from a video. To this end, we model the blurred appearance of a fast moving object in a generative fashion by parametrizing its 3D position, rotation, velocity, acceleration, bounces, shape, and texture over the duration of a predefined time window spanning multiple frames. Using differentiable rendering, we are able to estimate all parameters by minimizing the pixel-wise reprojection error to the input video via backpropagating through a rendering pipeline that accounts for motion blur by averaging the graphics output over short time intervals. For that purpose, we also estimate the camera exposure gap time within the same optimization. To account for abrupt motion changes like bounces, we model the motion trajectory as a piece-wise polynomial, and we are able to estimate the specific time of the bounce at sub-frame accuracy. Experiments on established benchmark datasets demonstrate that our method outperforms previous methods for fast moving object deblurring and 3D reconstruction.

We present IterMVS, a new data-driven method for high-resolution multi-view stereo. We propose a novel GRU-based estimator that encodes pixel-wise probability distributions of depth in its hidden state. Ingesting multi-scale matching information, our model refines these distributions over multiple iterations and infers depth and confidence. To extract the depth maps, we combine traditional classification and regression in a novel manner. We verify the efficiency and effectiveness of our method on DTU, Tanks&Temples and ETH3D. While being the most efficient method in both memory and run-time, our model achieves competitive performance on DTU and better generalization ability on Tanks&Temples as well as ETH3D than most state-of-the-art methods. Code is available at https://github.com/FangjinhuaWang/IterMVS.

Neural implicit representations have recently shown encouraging results in various domains, including promising progress in simultaneous localization and mapping (SLAM). Nevertheless, existing methods produce over- smoothed scene reconstructions and have difficulty scaling up to large scenes. These limitations are mainly due to their simple fully-connected network architecture that does not incorporate local information in the observations. In this paper, we present NICE-SLAM, a dense SLAM system that incorporates multi-level local information by introducing a hierarchical scene representation. Optimizing this representation with pre-trained geometric priors enables detailed reconstruction on large indoor scenes. Compared to recent neural implicit SLAM systems, our approach is more scalable, efficient, and robust. Experiments on five challenging datasets demonstrate competitive results of NICE-SLAM in both mapping and tracking quality. Project page: https://pengsongyou.github.io/nice-slam.

For robotic interaction in environments shared with other agents, access to volumetric and semantic maps of the scene is crucial. However, such environments are inevitably subject to long-term changes, which the map needs to account for. We thus propose panoptic multi-TSDFs as a novel representation for multi-resolution volumetric mapping in changing environments. By leveraging high-level information for 3D reconstruction, our proposed system allocates high resolution only where needed. Through reasoning on the object level, semantic consistency over time is achieved. This enables our method to maintain up-to-date reconstructions with high accuracy while improving coverage by incorporating previous data. We show in thorough experimental evaluation that our map can be efficiently constructed, maintained, and queried during online operation, and that the presented approach can operate robustly on real depth sensors using non-optimized panoptic segmentation as input.

While recent work has shown progress on extracting clothed 3D human avatars from a single image, video, or a set of 3D scans, several limitations remain. Most methods use a holistic representation to jointly model the body and clothing, which means that the clothing and body cannot be separated for applications like virtual try-on. Other methods separately model the body and clothing, but they require training from a large set of 3D clothed human meshes obtained from 3D/4D scanners or physics simulations. Our insight is that the body and clothing have different modeling requirements. While the body is well represented by a mesh-based parametric 3D model, implicit representations and neural radiance fields are better suited to capturing the large variety in shape and appearance present in clothing. Building on this insight, we propose SCARF (Segmented Clothed Avatar Radiance Field), a hybrid model combining a mesh-based body with a neural radiance field. Integrating the mesh into the volumetric rendering in combination with a differentiable rasterizer enables us to optimize SCARF directly from monocular videos, without any 3D supervision. The hybrid modeling enables SCARF to (i) animate the clothed body avatar by changing body poses (including hand articulation and facial expressions), (ii) synthesize novel views of the avatar, and (iii) transfer clothing between avatars in virtual try-on applications. We demonstrate that SCARF reconstructs clothing with higher visual quality than existing methods, that the clothing deforms with changing body pose and body shape, and that clothing can be successfully transferred between avatars of different subjects. The code and models are available at https://github.com/YadiraF/SCARF.

Representing scenes at the granularity of objects is a prerequisite for scene understanding and decision making. We propose PriSMONet, a novel approach based on Prior Shape knowledge for learning Multi-Object 3D scene decomposition and representations from single images. Our approach learns to decompose images of synthetic scenes with multiple objects on a planar surface into its constituent scene objects and to infer their 3D properties from a single view. A recurrent encoder regresses a latent representation of 3D shape, pose and texture of each object from an input RGB image. By differentiable rendering, we train our model to decompose scenes from RGB-D images in a self-supervised way. The 3D shapes are represented continuously in function-space as signed distance functions which we pre-train from example shapes in a supervised way. These shape priors provide weak supervision signals to better condition the challenging overall learning task. We evaluate the accuracy of our model in inferring 3D scene layout, demonstrate its generative capabilities, assess its generalization to real images, and point out benefits of the learned representation.

Visual localization enables autonomous vehicles to navigate in their surroundings and augmented reality applications to link virtual to real worlds. Practical visual localization approaches need to be robust to a wide variety of viewing conditions, including day-night changes, as well as weather and seasonal variations, while providing highly accurate six degree-of-freedom (6DOF) camera pose estimates. In this paper, we extend three publicly available datasets containing images captured under a wide variety of viewing conditions, but lacking camera pose information, with ground truth pose information, making evaluation of the impact of various factors on 6DOF camera pose estimation accuracy possible. We also discuss the performance of state-of-the-art localization approaches on these datasets. Additionally, we release around half of the poses for all conditions, and keep the remaining half private as a test set, in the hopes that this will stimulate research on long-term visual localization, learned local image features, and related research areas. Our datasets are available at visuallocalization.net, where we are also hosting a benchmarking server for automatic evaluation of results on the test set. The presented state-of-the-art results are to a large degree based on submissions to our server.

Building upon the recent progress in novel view synthesis, we propose its application to improve monocular depth estimation. In particular, we propose a novel training method split in three main steps. First, the prediction results of a monocular depth network are warped to an additional view point. Second, we apply an additional image synthesis network, which corrects and improves the quality of the warped RGB image. The output of this network is required to look as similar as possible to the ground-truth view by minimizing the pixel-wise RGB reconstruction error. Third, we reapply the same monocular depth estimation onto the synthesized second view point and ensure that the depth predictions are consistent with the associated ground truth depth. Experimental results prove that our method achieves state-of-the-art or comparable performance on the KITTI and NYU-Depth-v2 datasets with a lightweight and simple vanilla U-Net architecture.

3D representation and reconstruction of human bodies have been studied for a long time in computer vision. Traditional methods rely mostly on parametric statistical linear models, limiting the space of possible bodies to linear combinations. It is only recently that some approaches try to leverage neural implicit representations for human body modeling, and while demonstrating impressive results, they are either limited by representation capability or not physically meaningful and controllable. In this work, we propose a novel neural implicit representation for the human body, which is fully differentiable and optimizable with disentangled shape and pose latent spaces. Contrary to prior work, our representation is designed based on the kinematic model, which makes the representation controllable for tasks like pose animation, while simultaneously allowing the optimization of shape and pose for tasks like 3D fitting and pose tracking. Our model can be trained and fine-tuned directly on non-watertight raw data with well-designed losses. Experiments demonstrate the improved 3D reconstruction performance over SoTA approaches and show the applicability of our method to shape interpolation, model fitting, pose tracking, and motion retargeting.

Estimating 3D hand meshes from RGB images robustly is a highly desirable task, made challenging due to the numerous degrees of freedom, and issues such as self-similarity and occlusions. Previous methods generally either use parametric 3D hand models or follow a model-free approach. While the former can be considered more robust, e.g. to occlusions, they are less expressive. We propose a hybrid approach, utilizing a deep neural network and differential rendering based optimization to demonstrably achieve the best of both worlds. In addition, we explore Virtual Reality (VR) as an application. Most VR headsets are nowadays equipped with multiple cameras, which we can leverage by extending our method to the egocentric stereo domain. This extension proves to be more resilient to the above mentioned issues. Finally, as a use-case, we show that the improved image-model alignment can be used to acquire the user’s hand texture, which leads to a more realistic virtual hand representation.

Many computer vision systems require users to upload image features to the cloud for processing and storage. These features can be exploited to recover sensitive information about the scene or subjects, e.g., by reconstructing the appearance of the original image. To address this privacy concern, we propose a new privacy-preserving feature representation. The core idea of our work is to drop constraints from each feature descriptor by embedding it within an affine subspace containing the original feature as well as adversarial feature samples. Feature matching on the privacy-preserving representation is enabled based on the notion of subspace-to-subspace distance. We experimentally demonstrate the effectiveness of our method and its high practical relevance for the applications of visual localization and mapping as well as face authentication. Compared to the original features, our approach makes it significantly more difficult for an adversary to recover private information.

We propose an online multi-view depth prediction approach on posed video streams, where the scene geometry information computed in the previous time steps is propagated to the current time step in an efficient and geometrically plausible way. The backbone of our approach is a real-time capable, lightweight encoder-decoder that relies on cost volumes computed from pairs of images. We extend it by placing a ConvLSTM cell at the bottleneck layer, which compresses an arbitrary amount of past information in its states. The novelty lies in propagating the hidden state of the cell by accounting for the viewpoint changes between time steps. At a given time step, we warp the previous hidden state into the current camera plane using the previous depth prediction. Our extension brings only a small overhead of computation time and memory consumption, while improving the depth predictions significantly. As a result, we outperform the existing state-of-the-art multi-view stereo methods on most of the evaluated metrics in hundreds of indoor scenes while maintaining a real-time performance. Code available: https://github.com/ardaduz/deep-video-mvs

Recent works on localization and mapping from privacy preserving line features have made significant progress towards addressing the privacy concerns arising from cloud-based solutions in mixed reality and robotics. The requirement for calibrated cameras is a fundamental limitation for these approaches, which prevents their application in many crowd-sourced mapping scenarios. In this paper, we propose a solution to the uncalibrated privacy preserving localization and mapping problem. Our approach simultaneously recovers the intrinsic and extrinsic calibration of a camera from line-features only. This enables uncalibrated devices to both localize themselves within an existing map as well as contribute to the map, while preserving the privacy of the image contents. Furthermore, we also derive a solution to bootstrapping maps from scratch using only uncalibrated devices. Our approach provides comparable performance to the calibrated scenario and the privacy compromising alternatives based on traditional point features.

Compared to feature point detection and description, detecting and matching line segments offer additional challenges. Yet, line features represent a promising complement to points for multi-view tasks. Lines are indeed well-defined by the image gradient, frequently appear even in poorly textured areas and offer robust structural cues. We thus hereby introduce the first joint detection and description of line segments in a single deep network. Thanks to a self-supervised training, our method does not require any annotated line labels and can therefore generalize to any dataset. Our detector offers repeatable and accurate localization of line segments in images, departing from the wireframe parsing approach. Leveraging the recent progresses in descriptor learning, our proposed line descriptor is highly discriminative, while remaining robust to viewpoint changes and occlusions. We evaluate our approach against previous line detection and description methods on several multi-view datasets created with homographic warps as well as real-world viewpoint changes. Our full pipeline yields higher repeatability, localization accuracy and matching metrics, and thus represents a first step to bridge the gap with learned feature points methods. Code and trained weights are available at https://github.com/cvg/SOLD2.

Objects moving at high speed appear significantly blurred when captured with cameras. The blurry appearance is especially ambiguous when the object has complex shape or texture. In such cases, classical methods, or even humans, are unable to recover the object’s appearance and motion. We propose a method that, given a single image with its estimated background, outputs the object’s appearance and position in a series of sub-frames as if captured by a high-speed camera (i.e. temporal super-resolution). The proposed generative model embeds an image of the blurred object into a latent space representation, disentangles the background, and renders the sharp appearance. Inspired by the image formation model, we design novel self-supervised loss function terms that boost performance and show good generalization capabilities. The proposed DeFMO method is trained on a complex synthetic dataset, yet it performs well on real-world data from several datasets. DeFMO outperforms the state of the art and generates high-quality temporal super-resolution frames.

Camera pose estimation in known scenes is a 3D geometry task recently tackled by multiple learning algorithms. Many regress precise geometric quantities, like poses or 3D points, from an input image. This either fails to generalize to new viewpoints or ties the model parameters to a specific scene. In this paper, we go Back to the Feature: we argue that deep networks should focus on learning robust and invariant visual features, while the geometric estimation should be left to principled algorithms. We introduce PixLoc, a scene-agnostic neural network that estimates an accurate 6-DoF pose from an image and a 3D model. Our approach is based on the direct alignment of multiscale deep features, casting camera localization as metric learning. PixLoc learns strong data priors by end-to-end training from pixels to pose and exhibits exceptional generalization to new scenes by separating model parameters and scene geometry. The system can localize in large environments given coarse pose priors but also improve the accuracy of sparse feature matching by jointly refining keypoints and poses with little overhead. The code will be publicly available at github.com/cvg/pixloc.

We present PatchmatchNet, a novel and learnable cascade formulation of Patchmatch for high-resolution multi-view stereo. With high computation speed and low memory requirement, PatchmatchNet can process higher resolution imagery and is more suited to run on resource limited devices than competitors that employ 3D cost volume regularization. For the first time we introduce an iterative multi-scale Patchmatch in an end-to-end trainable architecture and improve the Patchmatch core algorithm with a novel and learned adaptive propagation and evaluation scheme for each iteration. Extensive experiments show a very competitive performance and generalization for our method on DTU, Tanks & Temples and ETH3D, but at a significantly higher efficiency than all existing top-performing models: at least two and a half times faster than state-of-the-art methods with twice less memory usage. Code is available at https://github.com/FangjinhuaWang/PatchmatchNet.

We present a novel online depth map fusion approach that learns depth map aggregation in a latent feature space. While previous fusion methods use an explicit scene representation like signed distance functions (SDFs), we propose a learned feature representation for the fusion. The key idea is a separation between the scene representation used for the fusion and the output scene representation, via an additional translator network. Our neural network architecture consists of two main parts: a depth and feature fusion sub-network, which is followed by a translator sub-network to produce the final surface representation (e.g. TSDF) for visualization or other tasks. Our approach is an online process, handles high noise levels, and is particularly able to deal with gross outliers common for photometric stereobased depth maps. Experiments on real and synthetic data demonstrate improved results compared to the state of the art, especially in challenging scenarios with large amounts of noise and outliers. The source code will be made available at https://github.com/weders/NeuralFusion.

We present a new pipeline for holistic 3D scene understanding from a single image, which could predict object shapes, object poses, and scene layout. As it is a highly ill-posed problem, existing methods usually suffer from inaccurate estimation of both shapes and layout especially for the cluttered scene due to the heavy occlusion between objects. We propose to utilize the latest deep implicit representation to solve this challenge. We not only propose an image-based local structured implicit network to improve the object shape estimation, but also refine the 3D object pose and scene layout via a novel implicit scene graph neural network that exploits the implicit local object features. A novel physical violation loss is also proposed to avoid incorrect context between objects. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods in terms of object shape, scene layout estimation, and 3D object detection.

Visual localization and mapping is the key technology underlying the majority of mixed reality and robotics systems. Most state-of-the-art approaches rely on local features to establish correspondences between images. In this paper, we present three novel scenarios for localization and mapping which require the continuous update of feature representations and the ability to match across different feature types. While localization and mapping is a fundamental computer vision problem, the traditional setup supposes the same local features are used throughout the evolution of a map. Thus, whenever the underlying features are changed, the whole process is repeated from scratch. However, this is typically impossible in practice, because raw images are often not stored and re-building the maps could lead to loss of the attached digital content. To overcome the limitations of current approaches, we present the first principled solution to cross-descriptor localization and mapping. Our data-driven approach is agnostic to the feature descriptor type, has low computational requirements, and scales linearly with the number of description algorithms. Extensive experiments demonstrate the effectiveness of our approach on state-of-the-art benchmarks for a variety of handcrafted and learned features.

We present a comprehensive framework for egocentric interaction recognition using markerless 3D annotations of two hands manipulating objects. To this end, we propose a method to create a unified dataset for egocentric 3D interaction recognition. Our method produces annotations of the 3D pose of two hands and the 6D pose of the manipulated objects, along with their interaction labels for each frame. Our dataset, called H2O (2 Hands and Objects), provides synchronized multi-view RGB-D images, interaction labels, object classes, ground-truth 3D poses for left & right hands, 6D object poses, ground-truth camera poses, object meshes and scene point clouds. To the best of our knowledge, this is the first benchmark that enables the study of first-person actions with the use of the pose of both left and right hands manipulating objects and presents an unprecedented level of detail for egocentric 3D interaction recognition. We further propose the method to predict interaction classes by estimating the 3D pose of two hands and the 6D pose of the manipulated objects, jointly from RGB images. Our method models both inter- and intra-dependencies between both hands and objects by learning the topology of a graph convolutional network that predicts interactions. We show that our method facilitated by this dataset establishes a strong baseline for joint hand-object pose estimation and achieves state-of-the-art accuracy for first person interaction recognition.

In this paper we consider the epipolar geometry between orthographic and perspective cameras. We generalize many of the classical results for the perspective essential matrix to this setting and derive novel minimal solvers, not only for the calibrated case, but also for partially calibrated and non-central camera setups. While orthographic cameras might seem exotic, they occur naturally in many applications. They can e.g. model 2D maps (such as floor plans), aerial/satellite photography and even approximate narrow field-of-view cameras (e.g. from telephoto lenses). In our experiments we highlight various applications of the developed theory and solvers, including Radar-Camera calibration and aligning Structure-from-Motion models to aerial or satellite images.

We aim at improving the computational efficiency of graph convolutional networks (GCNs) for learning on point clouds. The basic graph convolution that is composed of a K-nearest neighbor (KNN) search and a multilayer perceptron (MLP) is examined. By mathematically analyzing the operations there, two findings to improve the efficiency of GCNs are obtained. (1) The local geometric structure information of 3D representations propagates smoothly across the GCN that relies on KNN search to gather neighborhood features. This motivates the simplification of multiple KNN searches in GCNs. (2) Shuffling the order of graph feature gathering and an MLP leads to equivalent or similar composite operations. Based on those findings, we optimize the computational procedure in GCNs. A series of experiments show that the optimized networks have reduced computational complexity, decreased memory consumption, and accelerated inference speed while maintaining comparable accuracy for learning on point clouds.

We present a novel method for synthesizing both temporally and geometrically consistent street-view panoramic video from a single satellite image and camera trajectory. Existing cross-view synthesis approaches focus on images, while video synthesis in such a case has not yet received enough attention. For geometrical and temporal consistency, our approach explicitly creates a 3D point cloud representation of the scene and maintains dense 3D-2D correspondences across frames that reflect the geometric scene configuration inferred from the satellite view. As for synthesis in the 3D space, we implement a cascaded network architecture with two hourglass modules to generate point-wise coarse and fine features from semantics and per-class latent vectors, followed by projection to frames and an up-sampling module to obtain the final realistic video. By leveraging computed correspondences, the produced street-view video frames adhere to the 3D geometric scene structure and maintain temporal consistency. Qualitative and quantitative experiments demonstrate superior results compared to other state-of-the-art synthesis approaches that either lack temporal consistency or realistic appearance. To the best of our knowledge, our work is the first one to synthesize cross-view images to videos..

Finding local features that are repeatable across multiple views is a cornerstone of sparse 3D reconstruction. The classical image matching paradigm detects keypoints per-image once and for all, which can yield poorly-localized features and propagate large errors to the final geometry. In this paper, we refine two key steps of structure-from-motion by a direct alignment of low-level image information from multiple views: we first adjust the initial keypoint locations prior to any geometric estimation, and subsequently refine points and camera poses as a post-processing. This refinement is robust to large detection noise and appearance changes, as it optimizes a featuremetric error based on dense features predicted by a neural network. This significantly improves the accuracy of camera poses and scene geometry for a wide range of keypoint detectors, challenging viewing conditions, and off-the-shelf deep features. Our system easily scales to large image collections, enabling pixel-perfect crowd-sourced localization at scale. Our code is publicly available at github.com/cvg/pixel-perfect-sfm as an add-on to the popular SfM software COLMAP.

Motion blur is one of the major challenges remaining for visual odometry methods. In low-light conditions where longer exposure times are necessary, motion blur can appear even for relatively slow camera motions. In this paper we present a novel hybrid visual odometry pipeline with direct approach that explicitly models and estimates the camera’s local trajectory within the exposure time. This allows us to actively compensate for any motion blur that occurs due to the camera motion. In addition, we also contribute a novel benchmarking dataset for motion blur aware visual odometry. In experiments we show that by directly modeling the image formation process, we are able to improve robustness of the visual odometry, while keeping comparable accuracy as that for images without motion blur. Both the code and the datasets can be found from https://github.com/ethliup/MBA-VO.

We propose the first learning-based approach for fast moving objects detection. Such objects are highly blurred and move over large distances within one video frame. Fast moving objects are associated with a deblurring and matting problem, also called deblatting. We show that the separation of deblatting into consecutive matting and deblurring allows achieving real-time performance, i.e. an order of magnitude speed-up, and thus enabling new classes of application. The proposed method detects fast moving objects as a truncated distance function to the trajectory by learning from synthetic data. For the sharp appearance estimation and accurate trajectory estimation, we propose a matting and fitting network that estimates the blurred appearance without background, followed by an energy minimization based deblurring. The state-of-the-art methods are outperformed in terms of recall, precision, trajectory estimation, and sharp appearance reconstruction. Compared to other methods, such as deblatting, the inference is of several orders of magnitude faster and allows applications such as real-time fast moving object detection and retrieval in large video collections.

Recovering high-quality 3D human motion in complex scenes from monocular videos is important for many applications, ranging from AR/VR to robotics. However, capturing realistic human-scene interactions, while dealing with occlusions and partial views, is challenging; current approaches are still far from achieving compelling results. We address this problem by proposing LEMO: LEarning human MOtion priors for 4D human body capture. By leveraging the large-scale motion capture dataset AMASS [38], we introduce a novel motion smoothness prior, which strongly reduces the jitters exhibited by poses recovered over a sequence. Furthermore, to handle contacts and occlusions occurring frequently in body-scene interactions, we design a contact friction term and a contact-aware motion infiller obtained via per-instance self-supervised training. To prove the effectiveness of the proposed motion priors, we combine them into a novel pipeline for 4D human body capture in 3D scenes. With our pipeline, we demonstrate high-quality 4D human body capture, reconstructing smooth motions and physically plausible body-scene interactions. The code and data are available at https://sanweiliti.github.io/LEMO/LEMO.html.

In this work, we present a lightweight, tightly-coupled deep depth network and visual-inertial odometry (VIO) system, which can provide accurate state estimates and dense depth maps of the immediate surroundings. Leveraging the proposed lightweight Conditional Variational Autoencoder (CVAE) for depth inference and encoding, we provide the network with previously marginalized sparse features from VIO to increase the accuracy of initial depth prediction and generalization capability. The compact representation of dense depth, termed depth code, can be updated jointly with navigation states in a sliding window estimator in order to provide the dense local scene geometry. We additionally propose a novel method to obtain the CVAE’s Jacobian which is shown to be more than an order of magnitude faster than previous works, and we additionally leverage First-Estimate Jacobian (FEJ) to avoid recalculation. As opposed to previous works that rely on completely dense residuals, we propose to only provide sparse measurements to update the depth code and show through careful experimentation that our choice of sparse measurements and FEJs can still significantly improve the estimated depth maps. Our full system also exhibits state-of-the-art pose estimation accuracy, and we show that it can run in real-time with single-thread execution while utilizing GPU acceleration only for the network and code Jacobian.

In recent years, neural implicit representations gained popularity in 3D reconstruction due to their expressiveness and flexibility. However, the implicit nature of neural implicit representations results in slow inference times and requires careful initialization. In this paper, we revisit the classic yet ubiquitous point cloud representation and introduce a differentiable point-to-mesh layer using a differentiable formulation of Poisson Surface Reconstruction (PSR) which allows for a GPU-accelerated fast solution of the indicator function given an oriented point cloud. The differentiable PSR layer allows us to efficiently and differentiably bridge the explicit 3D point representation with the 3D mesh via the implicit indicator field, enabling end-to-end optimization of surface reconstruction metrics such as Chamfer distance. This duality between points and meshes hence allows us to represent shapes as oriented point clouds, which are explicit, lightweight and expressive. Compared to neural implicit representations, our Shape-As-Points (SAP) model is more interpretable, lightweight, and accelerates inference time by one order of magnitude. Compared to other explicit representations such as points, patches, and meshes, SAP produces topology-agnostic, watertight manifold surfaces. We demonstrate the effectiveness of SAP on the task of surface reconstruction from unoriented point clouds and learning-based reconstruction.

We address the novel task of jointly reconstructing the 3D shape, texture, and motion of an object from a single motion-blurred image. While previous approaches address the deblurring problem only in the 2D image domain, our proposed rigorous modeling of all object properties in the 3D domain enables the correct description of arbitrary object motion. This leads to significantly better image decomposition and sharper deblurring results. We model the observed appearance of a motion-blurred object as a combination of the background and a 3D object with constant translation and rotation. Our method minimizes a loss on reconstructing the input image via differentiable rendering with suitable regularizers. This enables estimating the textured 3D mesh of the blurred object with high fidelity. Our method substantially outperforms competing approaches on several benchmarks for fast moving objects deblurring. Qualitative results show that the reconstructed 3D mesh generates high-quality temporal super-resolution and novel views of the deblurred object.

We seek to predict the 6 degree-of-freedom (6DoF) pose of a query photograph with respect to a large indoor 3D map. The contributions of this work are three-fold. First, we develop a new large-scale visual localization method targeted for indoor spaces. The method proceeds along three steps: (i) efficient retrieval of candidate poses that scales to large-scale environments, (ii) pose estimation using dense matching rather than sparse local features to deal with weakly textured indoor scenes, and (iii) pose verification by virtual view synthesis that is robust to significant changes in viewpoint, scene layout, and occlusion. Second, we release a new dataset with reference 6DoF poses for large-scale indoor localization. Query photographs are captured by mobile phones at a different time than the reference 3D map, thus presenting a realistic indoor localization scenario. Third, we demonstrate that our method significantly outperforms current state-of-the-art indoor localization approaches on this new challenging data. Code and data are publicly available.

Accurate visual localization is a key technology for autonomous navigation. 3D structure-based methods employ 3D models of the scene to estimate the full 6 degree-of-freedom (DOF) pose of a camera very accurately. However, constructing (and extending) large-scale 3D models is still a significant challenge. In contrast, 2D image retrieval-based methods only require a database of geo-tagged images, which is trivial to construct and to maintain. They are often considered inaccurate since they only approximate the positions of the cameras. Yet, the exact camera pose can theoretically be recovered when enough relevant database images are retrieved. In this paper, we demonstrate experimentally that large-scale 3D models are not strictly necessary for accurate visual localization. We create reference poses for a large and challenging urban dataset. Using these poses, we show that combining image-based methods with local reconstructions results in a higher pose accuracy compared to state-of-the-art structure-based methods, albeight at higher run-time costs. We show that some of these run-time costs can be alleviated by exploiting known database image poses. Our results suggest that we might want to reconsider the need for large-scale 3D models in favor of more local models, but also that further research is necessary to accelerate the local reconstruction process.

Surface reconstruction from point clouds has been a well-studied research topic with applications in computer vision and computer graphics. Recently, several learning based methods were proposed for 3D shape representation through implicit functions which among others can be used for point cloud-based reconstruction. Although delivering compelling results for synthetic object datasets of overseeable size, they fail to represent larger scenes accurately, presumably due to the use of only one global latent code for encoding an entire scene or object. We propose to encode only parts of objects with features attached to unstructured point clouds. To this end we use a hierarchical feature map in 3D space, extracted from the input point clouds, with which local latent shape encodings can be queried at arbitrary positions. We use a permutohedral lattice to process the hierarchical feature maps sparsely and efficiently. This enables accurate and detailed point cloud-based reconstructions for large amounts of points in a time-efficient manner, showing good generalization capabilities across different datasets. Experiments on synthetic and real world datasets demonstrate the reconstruction capability of our method and compare favorably to state-of-the-art methods.

Recent progress in consumer hardware allowed for the collection of a large amount of animated point cloud data, which is on the one hand highly redundant and on the other hand incomplete. Our goal is to bridge this gap and find a low dimensional representation capable of approximation to a desired precision and completion of missing data. Model-less non-rigid 3D reconstruction algorithms, formulated as a linear factorization of observed point tracks into static shape component and dynamic pose, have been found insufficient to create suitable generative models, capable of generating new unobserved poses. This is due to the non-locality of the linear models, over-fitting to the non-causal correlations present in the data, which manifests in the reconstruction containing rigidly behaving not directly connected parts. In this paper, we propose a new method that can distinguish body parts and factorize the data into shape and pose purely using topological properties of the manifold-local deformations and neighborhoods. To obtain localized factorization, we formulate the deformation distance between two point tracks as the smallest deformation along the path between them. After embedding such distance in low dimensional space, a clustering of embedded data leads to close to rigid components, suitable as initialization for fitting a model-a skinned rigged mesh, used extensively in computer graphics. As both local deformations and neighborhoods of a point are local and can be estimated only from the part of the animation, the method can be used to recover unobserved data in each frame.

We present a novel 3D shape completion method that operates directly on unstructured point clouds, thus avoiding resource-intensive data structures like voxel grids. To this end, we introduce KAPLAN, a 3D point descriptor that aggregates local shape information via a series of 2D convolutions. The key idea is to project the points in a local neighborhood onto multiple planes with different orientations. In each of those planes, point properties like normals or point-to-plane distances are aggregated into a 2D grid and abstracted into a feature representation with an efficient 2D convolutional encoder. Since all planes are encoded jointly, the resulting representation nevertheless can capture their correlations and retains knowledge about the underlying 3D shape, without expensive 3D convolutions. Experiments on public datasets show that KAPLAN achieves state-of-the-art performance for 3D shape completion.

Decomposing scenes into reflectance and lighting is an important task for applications such as relighting, image matching or content creation. Advanced light transport effects like occlusion and indirect lighting are often ignored, leading to subpar decompositions in which the albedo needs to compensate for insufficiencies in the estimated shading. We show how to account for these advanced lighting effects by utilizing precomputed radiance transfer to estimate reflectance and lighting. Given the geometry of an object and one or multiple images, our method reconstructs the object's surface reflectance properties-such as its albedo and glossiness-as well as a colored lighting environment map. Evaluation on synthetic and real data shows that incorporation of indirect light leads to qualitatively and quantitatively improved results.

Most of the current scene flow methods choose to model scene flow as a per point translation vector without differentiating between static and dynamic components of 3D motion. In this work we present an alternative method for end-to-end scene flow learning by joint estimation of non-rigid residual flow and ego-motion flow for dynamic 3D scenes. We propose to learn the relative rigid transformation from a pair of point clouds followed by an iterative refinement. We then learn the non-rigid flow from transformed inputs with the deducted rigid part of the flow. Furthermore, we extend the supervised framework with self-supervisory signals based on the temporal consistency property of a point cloud sequence. Our solution allows both training in a supervised mode complemented by self-supervisory loss terms as well as training in a fully self-supervised mode. We demonstrate that decomposition of scene flow into non-rigid flow and ego-motion flow along with an introduction of the self-supervisory signals allowed us to outperform the current state-of-the-art supervised methods.

Modeling hand-object manipulations is essential for understanding how humans interact with their environment. While of practical importance, estimating the pose of hands and objects during interactions is challenging due to the large mutual occlusions that occur during manipulation. Recent efforts have been directed towards fully-supervised methods that require large amounts of labeled training samples. Collecting 3D ground-truth data for hand-object interactions, however, is costly, tedious, and error-prone. To overcome this challenge we present a method to leverage photometric consistency across time when annotations are only available for a sparse subset of frames in a video. Our model is trained end-to-end on color images to jointly reconstruct hands and objects in 3D by inferring their poses. Given our estimated reconstructions, we differentiably render the optical flow between pairs of adjacent images and use it within the network to warp one frame to another. We then apply a self-supervised photometric loss that relies on the visual consistency between nearby images. We achieve state-of-the-art results on 3D hand-object reconstruction benchmarks and demonstrate that our approach allows us to improve the pose estimation accuracy by leveraging information from neighboring frames in low-data regimes.

We present a novel network for rolling shutter effect correction. Our network takes two consecutive rolling shutter images and estimates the corresponding global shutter image of the latest frame. The dense displacement field from a rolling shutter image to its corresponding global shutter image is estimated via a motion estimation network. The learned feature representation of a rolling shutter image is then warped, via the displacement field, to its global shutter representation by a differentiable forward warping block. An image decoder recovers the global shutter image based on the warped feature representation. Our network can be trained end-to-end and only requires the global shutter image for supervision. Since there is no public dataset available, we also propose two large datasets: the Carla-RS dataset and the Fastec-RS dataset. Experimental results demonstrate that our network outperforms the state-of-the-art methods. We make both our code and datasets available at https://github.com/ethliup/DeepUnrollNet.

We propose a differentiable sphere tracing algorithm to bridge the gap between inverse graphics methods and the recently proposed deep learning based implicit signed distance function. Due to the nature of the implicit function, the rendering process requires tremendous function queries, which is particularly problematic when the function is represented as a neural network. We optimize both the forward and backward pass of our rendering layer to make it run efficiently with affordable memory consumption on a commodity graphics card. Our rendering method is fully differentiable such that losses can be directly computed on the rendered 2D observations, and the gradients can be propagated backward to optimize the 3D geometry. We show that our rendering method can effectively reconstruct accurate 3D shapes from various inputs, such as sparse depth and multi-view images, through inverse optimization. With the geometry based reasoning, our 3D shape prediction methods show excellent generalization capability and robustness against various noises.

We present a novel method for generating panoramic street-view images which are geometrically consistent with a given satellite image. Different from existing approaches that completely rely on a deep learning architecture to generalize cross-view image distributions, our approach explicitly loops in the geometric configuration of the ground objects based on the satellite views, such that the produced ground view synthesis preserves the geometric shape and the semantics of the scene. In particular, we propose a neural network with a geo-transformation layer that turns predicted ground-height values from the satellite view to a ground view while retaining the physical satellite-to-ground relation. Our results show that the synthesized image retains well-articulated and authentic geometric shapes, as well as texture richness of the street-view in various scenarios. Both qualitative and quantitative results demonstrate that our method compares favorably to other state-of-the-art approaches that lack geometric consistency.

Camera calibration is an essential first step in setting up 3D Computer Vision systems. Commonly used parametric camera models are limited to a few degrees of freedom and thus often do not optimally fit to complex real lens distortion. In contrast, generic camera models allow for very accurate calibration due to their flexibility. Despite this, they have seen little use in practice. In this paper, we argue that this should change. We propose a calibration pipeline for generic models that is fully automated, easy to use, and can act as a drop-in replacement for parametric calibration, with a focus on accuracy. We compare our results to parametric calibrations. Considering stereo depth estimation and camera pose estimation as examples, we show that the calibration error acts as a bias on the results. We thus argue that in contrast to current common practice, generic models should be preferred over parametric ones whenever possible. To facilitate this, we released our calibration pipeline at https://github.com/puzzlepaint/camera_calibration, making both easy-to-use and accurate camera calibration available to everyone.

Previous methods on estimating detailed human depth often require supervised training with ‘ground truth’ depth data. This paper presents a self-supervised method that can be trained on YouTube videos without known depth, which makes training data collection simple and improves the generalization of the learned network. The self-supervised learning is achieved by minimizing a photo-consistency loss, which is evaluated between a video frame and its neighboring frames warped according to the estimated depth and the 3D non-rigid motion of the human body. To solve this non-rigid motion, we first estimate a rough SMPL model at each video frame and compute the non-rigid body motion accordingly, which enables self-supervised learning on estimating the shape details. Experiments demonstrate that our method enjoys better generalization, and performs much better on data in the wild.

The efficient fusion of depth maps is a key part of most state-of-the-art 3D reconstruction methods. Besides requiring high accuracy, these depth fusion methods need to be scalable and real-time capable. To this end, we present a novel real-time capable machine learning-based method for depth map fusion. Similar to the seminal depth map fusion approach by Curless and Levoy, we only update a local group of voxels to ensure real-time capability. Instead of a simple linear fusion of depth information, we propose a neural network that predicts non-linear updates to better account for typical fusion errors. Our network is composed of a 2D depth routing network and a 3D depth fusion network which efficiently handle sensor-specific noise and outliers. This is especially useful for surface edges and thin objects for which the original approach suffers from thickening artifacts. Our method outperforms the traditional fusion approach and related learned approaches on both synthetic and real data. We demonstrate the performance of our method in reconstructing fine geometric details from noise and outlier contaminated data on various scenes.

In this paper, we present an omnidirectional localization and dense mapping system for a wide-baseline multiview stereo setup with ultra-wide field-of-view (FOV) fisheye cameras, which has a 360° coverage of stereo observations of the environment. For more practical and accurate reconstruction, we first introduce improved and light-weighted deep neural networks for the omnidirectional depth estimation, which are faster and more accurate than the existing networks. Second, we integrate our omnidirectional depth estimates into the visual odometry (VO) and add a loop closing module for global consistency. Using the estimated depth map, we reproject keypoints onto each other view, which leads to a better and more efficient feature matching process. Finally, we fuse the omnidirectional depth maps and the estimated rig poses into the truncated signed distance function (TSDF) volume to acquire a 3D map. We evaluate our method on synthetic datasets with ground-truth and real-world sequences of challenging environments, and the extensive experiments show that the proposed system generates excellent reconstruction results in both synthetic and real-world environments.

Visual localization is the problem of estimating a camera within a scene and a key technology for autonomous robots. State-of-the-art approaches for accurate visual localization use scene-specific representations, resulting in the overhead of constructing these models when applying the techniques to new scenes. Recently, learned approaches based on relative pose estimation have been proposed, carrying the promise of easily adapting to new scenes. However, they are currently significantly less accurate than state-of-the-art approaches. In this paper, we are interested in analyzing this behavior. To this end, we propose a novel framework for visual localization from relative poses. Using a classical feature-based approach within this framework, we show state-of-the-art performance. Replacing the classical approach with learned alternatives at various levels, we then identify the reasons for why deep learned approaches do not perform well. Based on our analysis, we make recommendations for future work.

Multi-sensor fusion of multi-modal measurements from commodity inertial, visual and LiDAR sensors to provide robust and accurate 6DOF pose estimation holds great potential in robotics and beyond. In this paper, building upon our prior work (i.e., LIC-Fusion), we develop a sliding-window filter based LiDAR-Inertial-Camera odometry with online spatiotemporal calibration (i.e., LIC-Fusion 2.0), which introduces a novel sliding-window plane-feature tracking for efficiently processing 3D LiDAR point clouds. In particular, after motion compensation for LiDAR points by leveraging IMU data, low-curvature planar points are extracted and tracked across the sliding window. A novel outlier rejection criteria is proposed in the plane-feature tracking for high quality data association. Only the tracked planar points belonging to the same plane will be used for plane initialization, which makes the plane extraction efficient and robust. Moreover, we perform the observability analysis for the IMU-LiDAR subsystem under consideration and report the degenerate cases for spatiotemporal calibration using plane features. While the estimation consistency and identified degenerate motions are validated in Monte-Carlo simulations, different real-world experiments are also conducted to show that the proposed LIC-Fusion 2.0 outperforms its predecessor and other state-of-the-art methods.

We address the problem of mesh reconstruction from live RGB-D video, assuming a calibrated camera and poses provided externally (e.g., by a SLAM system). In contrast to most existing approaches, we do not fuse depth measurements in a volume but in a dense surfel cloud. We asynchronously (re)triangulate the smoothed surfels to reconstruct a surface mesh. This novel approach enables to maintain a dense surface representation of the scene during SLAM which can quickly adapt to loop closures. This is possible by deforming the surfel cloud and asynchronously remeshing the surface where necessary. The surfel-based representation also naturally supports strongly varying scan resolution. In particular, it reconstructs colors at the input camera's resolution. Moreover, in contrast to many volumetric approaches, ours can reconstruct thin objects since objects do not need to enclose a volume. We demonstrate our approach in a number of experiments, showing that it produces reconstructions that are competitive with the state-of-the-art, and we discuss its advantages and limitations. The algorithm (excluding loop closure functionality) is available as open source at https://github.com/puzzlepaint/surfelmeshing.

We present a super-resolution method capable of creating a high-resolution texture map for a virtual 3D object from a set of lower-resolution images of that object. Our architecture unifies the concepts of (i) multi-view super-resolution based on the redundancy of overlapping views and (ii) single-view super-resolution based on a learned prior of high-resolution (HR) image structure. The principle of multi-view super-resolution is to invert the image formation process and recover the latent HR texture from multiple lower-resolution projections. We map that inverse problem into a block of suitably designed neural network layers, and combine it with a standard encoder-decoder network for learned single-image super-resolution. Wiring the image formation model into the network avoids having to learn perspective mapping from textures to images, and elegantly handles a varying number of input views. Experiments demonstrate that the combination of multi-view observations and learned prior yields improved texture maps.

Localizing an image w.r.t. a 3D scene model represents a core task for many computer vision applications. An increasing number of real-world applications of visual localization on mobile devices, e.g., Augmented Reality or autonomous robots such as drones or self-driving cars, demand localization approaches to minimize storage and bandwidth requirements. Compressing the 3D models used for localization thus becomes a practical necessity. In this work, we introduce a new hybrid compression algorithm that uses a given memory limit in a more effective way. Rather than treating all 3D points equally, it represents a small set of points with full appearance information and an additional, larger set of points with compressed information. This enables our approach to obtain a more complete scene representation without increasing the memory requirements, leading to a superior performance compared to previous compression schemes. As part of our contribution, we show how to handle ambiguous matches arising from point compression during RANSAC. Besides outperforming previous compression techniques in terms of pose accuracy under the same memory constraints, our compression scheme itself is also more efficient. Furthermore, the localization rates and accuracy obtained with our approach are comparable to state-of-the-art feature-based methods, while using a small fraction of the memory.

In this work we address the problem of finding reliable pixel-level correspondences under difficult imaging conditions. We propose an approach where a single convolutional neural network plays a dual role: It is simultaneously a dense feature descriptor and a feature detector. By postponing the detection to a later stage, the obtained keypoints are more stable than their traditional counterparts based on early detection of low-level structures. We show that this model can be trained using pixel correspondences extracted from readily available large-scale SfM reconstructions, without any further annotations. The proposed method obtains state-of-the-art performance on both the difficult Aachen Day-Night localization dataset and the InLoc indoor localization benchmark, as well as competitive performance on other benchmarks for image matching and 3D reconstruction.

In this paper, we present a method to utilize 2D-2D point matches between images taken during different image conditions to train a convolutional neural network for semantic segmentation. Enforcing label consistency across the matches makes the final segmentation algorithm robust to seasonal changes. We describe how these 2D-2D matches can be generated with little human interaction by geometrically matching points from 3D models built from images. Two cross-season correspondence datasets are created providing 2D-2D matches across seasonal changes as well as from day to night. The datasets are made publicly available to facilitate further research. We show that adding the correspondences as extra supervision during training improves the segmentation performance of the convolutional neural network, making it more robust to seasonal changes and weather conditions.

We tackle the problem of retrieving high-resolution (HR) texture maps of objects that are captured from multiple view points. In the multi-view case, model-based super-resolution (SR) methods have been recently proved to recover high quality texture maps. On the other hand, the advent of deep learning-based methods has already a significant impact on the problem of video and image SR. Yet, a deep learning-based approach to super-resolve the appearance of 3D objects is still missing. The main limitation of exploiting the power of deep learning techniques in the multi-view case is the lack of data. We introduce a 3D appearance SR (3DASR) dataset based on the existing ETH3D [42], SyB3R [31], MiddleBury, and our Collection of 3D scenes from TUM [21], Fountain [51] and Relief [53]. We provide the high- and low-resolution texture maps, the 3D geometric model, images and projection matrices. We exploit the power of 2D learning-based SR methods and design networks suitable for the 3D multi-view case. We incorporate the geometric information by introducing normal maps and further improve the learning process. Experimental results demonstrate that our proposed networks successfully incorporate the 3D geometric information and super-resolve the texture maps.

In this paper, we propose a deep learning architecture that produces accurate dense depth for the outdoor scene from a single color image and a sparse depth. Inspired by the indoor depth completion, our network estimates surface normals as the intermediate representation to produce dense depth, and can be trained end-to-end. With a modified encoder-decoder structure, our network effectively fuses the dense color image and the sparse LiDAR depth. To address outdoor specific challenges, our network predicts a confidence mask to handle mixed LiDAR signals near foreground boundaries due to occlusion, and combines estimates from the color image and surface normals with learned attention maps to improve the depth accuracy especially for distant areas. Extensive experiments demonstrate that our model improves upon the state-of-the-art performance on KITTI depth completion benchmark. Ablation study shows the positive impact of each model components to the final performance, and comprehensive analysis shows that our model generalizes well to the input with higher sparsity or from indoor scenes.

Visual localization is the task of accurate camera pose estimation in a known scene. It is a key problem in computer vision and robotics, with applications including self-driving cars, Structure-from-Motion, SLAM, and Mixed Reality. Traditionally, the localization problem has been tackled using 3D geometry. Recently, end-to-end approaches based on convolutional neural networks have become popular. These methods learn to directly regress the camera pose from an input image. However, they do not achieve the same level of pose accuracy as 3D structure-based methods. To understand this behavior, we develop a theoretical model for camera pose regression. We use our model to predict failure cases for pose regression techniques and verify our predictions through experiments. We furthermore use our model to show that pose regression is more closely related to pose approximation via image retrieval than to accurate pose estimation via 3D structure. A key result is that current approaches do not consistently outperform a handcrafted image retrieval baseline. This clearly shows that additional research is needed before pose regression algorithms are ready to compete with structure-based methods.

A key component of Simultaneous Localization and Mapping (SLAM) systems is the joint optimization of the estimated 3D map and camera trajectory. Bundle adjustment (BA) is the gold standard for this. Due to the large number of variables in dense RGB-D SLAM, previous work has focused on approximating BA. In contrast, in this paper we present a novel, fast direct BA formulation which we implement in a real-time dense RGB-D SLAM algorithm. In addition, we show that direct RGB-D SLAM systems are highly sensitive to rolling shutter, RGB and depth sensor synchronization, and calibration errors. In order to facilitate state-of-the-art research on direct RGB-D SLAM, we propose a novel, well-calibrated benchmark for this task that uses synchronized global shutter RGB and depth cameras. It includes a training set, a test set without public ground truth, and an online evaluation service. We observe that the ranking of methods changes on this dataset compared to existing ones, and our proposed algorithm outperforms all other evaluated SLAM methods. Our benchmark and our open source SLAM algorithm are available at: www.eth3d.net

Image-based localization is a core component of many augmented/mixed reality (AR/MR) and autonomous robotic systems. Current localization systems rely on the persistent storage of 3D point clouds of the scene to enable camera pose estimation, but such data reveals potentially sensitive scene information. This gives rise to significant privacy risks, especially as for many applications 3D mapping is a background process that the user might not be fully aware of. We pose the following question: How can we avoid disclosing confidential information about the captured 3D scene, and yet allow reliable camera pose estimation? This paper proposes the first solution to what we call privacy preserving image-based localization. The key idea of our approach is to lift the map representation from a 3D point cloud to a 3D line cloud. This novel representation obfuscates the underlying scene geometry while providing sufficient geometric constraints to enable robust and accurate 6-DOF camera pose estimation. Extensive experiments on several datasets and localization scenarios underline the high practical relevance of our proposed approach.

We present a unified framework for understanding 3D hand and object interactions in raw image sequences from egocentric RGB cameras. Given a single RGB image, our model jointly estimates the 3D hand and object poses, models their interactions, and recognizes the object and action classes with a single feed-forward pass through a neural network. We propose a single architecture that does not rely on external detection algorithms but rather is trained end-to-end on single images. We further merge and propagate information in the temporal domain to infer interactions between hand and object trajectories and recognize actions. The complete model takes as input a sequence of frames and outputs per-frame 3D hand and object pose predictions along with the estimates of object and action categories for the entire sequence. We demonstrate state-of-the-art performance of our algorithm even in comparison to the approaches that work on depth data and ground-truth annotations.

In this paper we consider the problem of relative pose estimation from two images with per-pixel polarimetric information. Using these additional measurements we derive a simple minimal solver for the essential matrix which only requires two point correspondences. The polarization constraints allow us to pointwise recover the 3D surface normal up to a two-fold ambiguity for the diffuse reflection. Since this ambiguity exists per point, there is a combinatorial explosion of possibilities. However, since our solver only requires two point correspondences, we only need to consider 16 configurations when solving for the relative pose. Once the relative orientation is recovered, we show that it is trivial to resolve the ambiguity for the remaining points. For robustness, we also propose a joint optimization between the relative pose and the refractive index to handle the refractive distortion. In experiments, on both synthetic and real data, we demonstrate that by leveraging the additional information available from polarization cameras, we can improve over classical methods which only rely on the 2D-point locations to estimate the geometry. Finally, we demonstrate the practical applicability of our approach by integrating it into a state-of-the-art global Structure-from-Motion pipeline.

We propose a novel method for instance label segmentation of dense 3D voxel grids. We target volumetric scene representations, which have been acquired with depth sensors or multi-view stereo methods and which have been processed with semantic 3D reconstruction or scene completion methods. The main task is to learn shape information about individual object instances in order to accurately separate them, including connected and incompletely scanned objects. We solve the 3D instance-labeling problem with a multi-task learning strategy. The first goal is to learn an abstract feature embedding, which groups voxels with the same instance label close to each other while separating clusters with different instance labels from each other. The second goal is to learn instance information by densely estimating directional information of the instance's center of mass for each voxel. This is particularly useful to find instance boundaries in the clustering post-processing step, as well as, for scoring the segmentation quality for the first goal. Both synthetic and real-world experiments demonstrate the viability and merits of our approach. In fact, it achieves state-of-the-art performance on the ScanNet 3D instance segmentation benchmark.

To model radial distortion there are two main approaches; either the image points are undistorted such that they correspond to pinhole projections, or the pinhole projections are distorted such that they align with the image measurements. Depending on the application, either of the two approaches can be more suitable. For example, distortion models are commonly used in Structure-from-Motion since they simplify measuring the reprojection error in images. Surprisingly, all previous minimal solvers for pose estimation with radial distortion use undistortion models. In this paper we aim to fill this gap in the literature by proposing the first minimal solvers which can jointly estimate distortion models together with camera pose. We present a general approach which can handle rational models of arbitrary degree for both distortion and undistortion.

Augmented/mixed reality and robotic applications are increasingly relying on cloud-based localization services, which require users to upload query images to perform camera pose estimation on a server. This raises significant privacy concerns when consumers use such services in their homes or in confidential industrial settings. Even if only image features are uploaded, the privacy concerns remain as the images can be reconstructed fairly well from feature locations and descriptors. We propose to conceal the content of the query images from an adversary on the server or a man-in-the-middle intruder. The key insight is to replace the 2D image feature points in the query image with randomly oriented 2D lines passing through their original 2D positions. It will be shown that this feature representation hides the image contents, and thereby protects user privacy, yet still provides sufficient geometric constraints to enable robust and accurate 6-DOF camera pose estimation from feature correspondences. Our proposed method can handle single- and multi-image queries as well as exploit additional information about known structure, gravity, and scale. Numerous experiments demonstrate the high practical relevance of our approach.

Visual localization is a key step in many robotics pipelines, allowing the robot to (approximately) determine its position and orientation in the world. An efficient and scalable approach to visual localization is to use image retrieval techniques. These approaches identify the image most similar to a query photo in a database of geo-tagged images and approximate the query's pose via the pose of the retrieved database image. However, image retrieval across drastically different illumination conditions, e.g. day and night, is still a problem with unsatisfactory results, even in this age of powerful neural models. This is due to a lack of a suitably diverse dataset with true correspondences to perform end-to-end learning. A recent class of neural models allows for realistic translation of images among visual domains with relatively little training data and, most importantly, without ground-truth pairings.In this paper, we explore the task of accurately localizing images captured from two traversals of the same area in both day and night. We propose ToDayGAN - a modified image-translation model to alter nighttime driving images to a more useful daytime representation. We then compare the daytime and translated night images to obtain a pose estimate for the night image using the known 6-DOF position of the closest day image. Our approach improves localization performance by over 250% compared the current state-of-the-art, in the context of standard metrics in multiple categories.

We present a real-time dense geometric mapping algorithm for large-scale environments. Unlike existing methods which use pinhole cameras, our implementation is based on fisheye cameras whose large field of view benefits various computer vision applications for self-driving vehicles such as visual-inertial odometry, visual localization, and object detection. Our algorithm runs on in-vehicle PCs at approximately 15 Hz, enabling vision-only 3D scene perception for self-driving vehicles. For each synchronized set of images captured by multiple cameras, we first compute a depth map for a reference camera using plane-sweeping stereo. To maintain both accuracy and efficiency, while accounting for the fact that fisheye images have a lower angular resolution, we recover the depths using multiple image resolutions. We adopt the fast object detection framework, YOLOv3, to remove potentially dynamic objects. At the end of the pipeline, we fuse the fisheye depth images into the truncated signed distance function (TSDF) volume to obtain a 3D map. We evaluate our method on large-scale urban datasets, and results show that our method works well in complex dynamic environments.

Visual localization, i.e., determining the position and orientation of a vehicle with respect to a map, is a key problem in autonomous driving. We present a multi-camera visual inertial localization algorithm for large scale environments. To efficiently and effectively match features against a pre-built global 3D map, we propose a prioritized feature matching scheme for multi-camera systems. In contrast to existing works, designed for monocular cameras, we (1) tailor the prioritization function to the multi-camera setup and (2) run feature matching and pose estimation in parallel. This significantly accelerates the matching and pose estimation stages and allows us to dynamically adapt the matching efforts based on the surrounding environment. In addition, we show how pose priors can be integrated into the localization system to increase efficiency and robustness. Finally, we extend our algorithm by fusing the absolute pose estimates with motion estimates from a multi-camera visual inertial odometry pipeline (VIO). This results in a system that provides reliable and drift-less pose estimation. Extensive experiments show that our localization runs fast and robust under varying conditions, and that our extended algorithm enables reliable real-time pose estimation.

Project AutoVision aims to develop localization and 3D scene perception capabilities for a self-driving vehicle. Such capabilities will enable autonomous navigation in urban and rural environments, in day and night, and with cameras as the only exteroceptive sensors. The sensor suite employs many cameras for both 360-degree coverage and accurate multi-view stereo; the use of low-cost cameras keeps the cost of this sensor suite to a minimum. In addition, the project seeks to extend the operating envelope to include GNSS-less conditions which are typical for environments with tall buildings, foliage, and tunnels. Emphasis is placed on leveraging multi-view geometry and deep learning to enable the vehicle to localize and perceive in 3D space. This paper presents an overview of the project, and describes the sensor suite and current progress in the areas of calibration, localization, and perception.

Visual-Inertial Odometry (VIO) algorithms typically rely on a point cloud representation of the scene that does not model the topology of the environment. A 3D mesh instead offers a richer, yet lightweight, model. Nevertheless, building a 3D mesh out of the sparse and noisy 3D landmarks triangulated by a VIO algorithm often results in a mesh that does not fit the real scene. In order to regularize the mesh, previous approaches decouple state estimation from the 3D mesh regularization step, and either limit the 3D mesh to the current frame [1], [2] or let the mesh grow indefinitely [3], [4]. We propose instead to tightly couple mesh regularization and state estimation by detecting and enforcing structural regularities in a novel factor-graph formulation. We also propose to incrementally build the mesh by restricting its extent to the time-horizon of the VIO optimization; the resulting 3D mesh covers a larger portion of the scene than a per-frame approach while its memory usage and computational complexity remain bounded. We show that our approach successfully regularizes the mesh, while improving localization accuracy, when structural regularities are present, and remains operational in scenes without regularities.

When we take photos through glass windows or doors, the transmitted background scene is often blended with undesirable reflection. Separating two layers apart to enhance the image quality is of vital importance for both human and machine perception. In this paper, we propose to exploit physical constraints from a pair of unpolarized and polarized images to separate reflection and transmission layers. Due to the simplified capturing setup, the system becomes more underdetermined compared with existing polarization based solutions that take three or more images as input. We propose to solve semireflector orientation estimation first to make the physical image formation well-posed and then learn to reliably separate two layers using a refinement network with gradient loss. Quantitative and qualitative experimental results show our approach performs favorably over existing polarization and single image based solutions.

This paper addresses the challenge of dense pixel correspondence estimation between two images. This problem is closely related to optical flow estimation task where ConvNets (CNNs) have recently achieved significant progress. While optical flow methods produce very accurate results for the small pixel translation and limited appearance variation scenarios, they hardly deal with the strong geometric transformations that we consider in this work. In this paper, we propose a coarse-to-fine CNN-based framework that can leverage the advantages of optical flow approaches and extend them to the case of large transformations providing dense and subpixel accurate estimates. It is trained on synthetic transformations and demonstrates very good performance to unseen, realistic, data. Further, we apply our method to the problem of relative camera pose estimation and demonstrate that the model outperforms existing dense approaches.

Abstract This work presents and evaluates a novel compact scene representation based on Stixels that infers geometric and semantic information. Our approach overcomes the previous rather restrictive geometric assumptions for Stixels by introducing a novel depth model to account for non-flat roads and slanted objects. Both semantic and depth cues are used jointly to infer the scene representation in a sound global energy minimization formulation. Furthermore, a novel approximation scheme is introduced in order to significantly reduce the computational complexity of the Stixel algorithm, and then achieve real-time computation capabilities. The idea is to first perform an over-segmentation of the image, discarding the unlikely Stixel cuts, and apply the algorithm only on the remaining Stixel cuts. This work presents a novel over-segmentation strategy based on a fully convolutional network, which outperforms an approach based on using local extrema of the disparity map. We evaluate the proposed methods in terms of semantic and geometric accuracy as well as run-time on four publicly available benchmark datasets. Our approach maintains accuracy on flat road scene datasets while improving substantially on a novel non-flat road dataset.

In this paper, we aim to solve the pose estimation problem of calibrated pinhole and generalized cameras w.r.t. a Structure-from-Motion (SfM) model by leveraging both 2D-3D correspondences as well as 2D-2D correspondences. Traditional approaches either focus on the use of 2D-3D matches, known as structure-based pose estimation or solely on 2D-2D matches (structure-less pose estimation). Absolute pose approaches are limited in their performance by the quality of the 3D point triangulations as well as the completeness of the 3D model. Relative pose approaches, on the other hand, while being more accurate, also tend to be far more computationally costly and often return dozens of possible solutions. This work aims to bridge the gap between these two paradigms. We propose a new RANSAC-based approach that automatically chooses the best type of solver to use at each iteration in a data-driven way. The solvers chosen by our RANSAC can range from pure structure-based or structure-less solvers, to any possible combination of hybrid solvers (i.e. using both types of matches) in between. A number of these new hybrid minimal solvers are also presented in this paper. Both synthetic and real data experiments show our approach to be as accurate as structure-less approaches, while staying close to the efficiency of structure-based methods.

Image-based 3D reconstruction for Internet photo collections has become a robust technology to produce impressive virtual representations of real-world scenes. However, several fundamental challenges remain for Structure-from-Motion (SfM) pipelines, namely: the placement and reconstruction of transient objects only observed in single views, estimating the absolute scale of the scene, and (suprisingly often) recovering ground surfaces in the scene. We propose a method to jointly address these remaining open problems of SfM. In particular, we focus on detecting people in individual images and accurately placing them into an existing 3D model. As part of this placement, our method also estimates the absolute scale of the scene from object semantics, which in this case constitutes the height distribution of the population. Further, we obtain a smooth approximation of the ground surface and recover the gravity vector of the scene directly from the individual person detections. We demonstrate the results of our approach on a number of unordered Internet photo collections, and we quantitatively evaluate the obtained absolute scene scales.

Visual localization enables autonomous vehicles to navigate in their surroundings and augmented reality applications to link virtual to real worlds. Practical visual localization approaches need to be robust to a wide variety of viewing condition, including day-night changes, as well as weather and seasonal variations, while providing highly accurate 6 degree-of-freedom (6DOF) camera pose estimates. In this paper, we introduce the first benchmark datasets specifically designed for analyzing the impact of such factors on visual localization. Using carefully created ground truth poses for query images taken under a wide variety of conditions, we evaluate the impact of various factors on 6DOF camera pose estimation accuracy through extensive experiments with state-of-the-art localization approaches. Based on our results, we draw conclusions about the difficulty of different conditions, showing that long-term localization is far from solved, and propose promising avenues for future work, including sequence-based localization approaches and the need for better local features. Our benchmark is available at visuallocalization.net.

Robust visual localization under a wide range of viewing conditions is a fundamental problem in computer vision. Handling the difficult cases of this problem is not only very challenging but also of high practical relevance, e.g., in the context of life-long localization for augmented reality or autonomous robots. In this paper, we propose a novel approach based on a joint 3D geometric and semantic understanding of the world, enabling it to succeed under conditions where previous approaches failed. Our method leverages a novel generative model for descriptor learning, trained on semantic scene completion as an auxiliary task. The resulting 3D descriptors are robust to missing observations by encoding high-level 3D geometric and semantic information. Experiments on several challenging large-scale localization datasets demonstrate reliable localization under extreme viewpoint, illumination, and geometry changes.

We propose a novel method for the geometric registration of semantically labeled regions. We approximate semantic regions by ellipsoids, and leverage their convexity to formulate the correspondence search effectively as a constrained optimization problem that maximizes the number of matched regions, and which we solve globally optimal in a Branch-and-Bound fashion. To this end, we derive suitable linear matrix inequality constraints which describe ellipsoid-to-ellipsoid assignment conditions. Our approach is robust to large percentages of outliers and thus applicable to difficult correspondence search problems. In multiple experiments we demonstrate the flexibility and robustness of our approach on a number of challenging vision problems.

We seek to predict the 6 degree-of-freedom (6DoF) pose of a query photograph with respect to a large indoor 3D map. The contributions of this work are three-fold. First, we develop a new large-scale visual localization method targeted for indoor environments. The method proceeds along three steps: (i) efficient retrieval of candidate poses that ensures scalability to large-scale environments, (ii) pose estimation using dense matching rather than local features to deal with texture less indoor scenes, and (iii) pose verification by virtual view synthesis to cope with significant changes in viewpoint, scene layout, and occluders. Second, we collect a new dataset with reference 6DoF poses for large-scale indoor localization. Query photographs are captured by mobile phones at a different time than the reference 3D map, thus presenting a realistic indoor localization scenario. Third, we demonstrate that our method significantly outperforms current state-of-the-art indoor localization approaches on this new challenging data.

We present a stereo-based dense mapping algorithm for large-scale dynamic urban environments. In contrast to other existing methods, we simultaneously reconstruct the static background, the moving objects, and the potentially moving but currently stationary objects separately, which is desirable for high-level mobile robotic tasks such as path planning in crowded environments. We use both instance-aware semantic segmentation and sparse scene flow to classify objects as either background, moving, or potentially moving, thereby ensuring that the system is able to model objects with the potential to transition from static to dynamic, such as parked cars. Given camera poses estimated from visual odometry, both the background and the (potentially) moving objects are reconstructed separately by fusing the depth maps computed from the stereo input. In addition to visual odometry, sparse scene flow is also used to estimate the 3D motions of the detected moving objects, in order to reconstruct them accurately. A map pruning technique is further developed to improve reconstruction accuracy and reduce memory consumption, leading to increased scalability. We evaluate our system thoroughly on the well-known KITTI dataset. Our system is capable of running on a PC at approximately 2.5Hz, with the primary bottleneck being the instance-aware semantic segmentation, which is a limitation we hope to address in future work. The source code is available from the project Websitea.

We present a visual odometry (VO) algorithm for a multi-camera system and robust operation in challenging environments. Our algorithm consists of a pose tracker and a local mapper. The tracker estimates the current pose by minimizing photometric errors between the most recent keyframe and the current frame. The mapper initializes the depths of all sampled feature points using plane-sweeping stereo. To reduce pose drift, a sliding window optimizer is used to refine poses and structure jointly. Our formulation is flexible enough to support an arbitrary number of stereo cameras. We evaluate our algorithm thoroughly on five datasets. The datasets were captured in different conditions: daytime, night-time with near-infrared (NIR) illumination and nighttime without NIR illumination. Experimental results show that a multi-camera setup makes the VO more robust to challenging environments, especially night-time conditions, in which a single stereo configuration fails easily due to the lack of features.

Accurate representation of soft transitions between image regions is essential for high-quality image editing and compositing. Current techniques for generating such representations depend heavily on interaction by a skilled visual artist, as creating such accurate object selections is a tedious task. In this work, we introduce semantic soft segments , a set of layers that correspond to semantically meaningful regions in an image with accurate soft transitions between different objects. We approach this problem from a spectral segmentation angle and propose a graph structure that embeds texture and color features from the image as well as higher-level semantic information generated by a neural network. The soft segments are generated via eigendecomposition of the carefully constructed Laplacian matrix fully automatically. We demonstrate that otherwise complex image editing tasks can be done with little effort using semantic soft segments.

Many geometric quantities can be computed efficiently for convex meshes. For general meshes, methods for approximate convex decomposition have been developed that decompose a static, non-convex object into a small set of approximately convex parts. The convex hulls of those parts can then be used as a piecewise convex approximation to the original mesh. While previous work was only concerned with static meshes, we present a method for decomposing animated 3D meshes into temporally coherent approximately convex parts. Given a mesh and several training frames---that is, different spatial configurations of its vertices---we precompute an approximate convex decomposition that is independent of any specific frame. Such a decomposition can be transferred in real-time to novel, unseen frames. We apply our method to a variety of pre-animated meshes as well as a 3D character interactively controlled by a user's body pose. We further demonstrate that our method enables real-time physics simulations to interact with animated meshes.

Symmetries and repetitions are common and valuable features in urban scenes. We propose to leverage such regularity information in an efficient optimization scheme in order to segment a rectified image of a facade into semantic categories. Our method retrieves a parsing which respects common architectural constraints as well as detected repetitive structures and edge information. Additionally, the use of symmetry information allows us to efficiently deal with large occluded areas and to recover plausible facade images with a minimum of occlusions. Our approach yields state-of-the-art accuracy on datasets with challenging occlusions. Competitive works either fully fail to deal with large occlusions or they are an order of magnitude slower than our approach.

We present a novel, purely affinity-based natural image matting algorithm. Our method relies on carefully defined pixel-to-pixel connections that enable effective use of information available in the image and the trimap. We control the information flow from the known-opacity regions into the unknown region, as well as within the unknown region itself, by utilizing multiple definitions of pixel affinities. This way we achieve significant improvements on matte quality near challenging regions of the foreground object. Among other forms of information flow, we introduce color-mixture flow, which builds upon local linear embedding and effectively encapsulates the relation between different pixel opacities. Our resulting novel linear system formulation can be solved in closed-form and is robust against several fundamental challenges in natural matting such as holes and remote intricate structures. While our method is primarily designed as a standalone natural matting tool, we show that it can also be used for regularizing mattes obtained by various sampling-based methods. Our evaluation using the public alpha matting benchmark suggests a significant performance improvement over the state-of-the-art.

Localizing a query image against a 3D model at large scale is a hard problem, since 2D-3D matches become more and more ambiguous as the model size increases. This creates a need for pose estimation strategies that can handle very low inlier ratios. In this paper, we draw new insights on the geometric information available from the 2D-3D matching process. As modern descriptors are not invariant against large variations in viewpoint, we are able to find the rays in space used to triangulate a given point that are closest to a query descriptor. It is well known that two correspondences constrain the camera to lie on the surface of a torus. Adding the knowledge of direction of triangulation, we are able to approximate the position of the camera from two matches alone. We derive a geometric solver that can compute this position in under 1 microsecond. Using this solver, we propose a simple yet powerful outlier filter which scales quadratically in the number of matches. We validate the accuracy of our solver and demonstrate the usefulness of our method in real world settings.

We present a method for the fast 3D face reconstruction of people wearing glasses. Our method explicitly and robustly models the case in which a face to be reconstructed is partially occluded by glasses. We propose a simple and generic model for glasses that copes with a wide variety of different shapes, colors and styles, without the need for any database or learning. Our algorithm is simple, fast and requires only small amounts of both memory and runtime resources, allowing for a fast interactive 3D reconstruction on commodity mobile phones. The thorough evaluation of our approach on synthetic and real data demonstrates superior reconstruction results due to the explicit modeling of glasses.

Accurate visual localization is a key technology for autonomous navigation. 3D structure-based methods employ 3D models of the scene to estimate the full 6DOF pose of a camera very accurately. However, constructing (and extending) large-scale 3D models is still a significant challenge. In contrast, 2D image retrieval-based methods only require a database of geo-tagged images, which is trivial to construct and to maintain. They are often considered inaccurate since they only approximate the positions of the cameras. Yet, the exact camera pose can theoretically be recovered when enough relevant database images are retrieved. In this paper, we demonstrate experimentally that large-scale 3D models are not strictly necessary for accurate visual localization. We create reference poses for a large and challenging urban dataset. Using these poses, we show that combining image-based methods with local reconstructions results in a pose accuracy similar to the state-of-the-art structure-based methods. Our results suggest that we might want to reconsider the current approach for accurate large-scale localization.

Several machine learning tasks require to represent the data using only a sparse set of interest points. An ideal detector is able to find the corresponding interest points even if the data undergo a transformation typical for a given domain. Since the task is of high practical interest in computer vision, many hand-crafted solutions were proposed. In this paper, we ask a fundamental question: can we learn such detectors from scratch? Since it is often unclear what points are interesting, human labelling cannot be used to find a truly unbiased solution. Therefore, the task requires an unsupervised formulation. We are the first to propose such a formulation: training a neural network to rank points in a transformation-invariant manner. Interest points are then extracted from the top/bottom quantiles of this ranking. We validate our approach on two tasks: standard RGB image interest point detection and challenging cross-modal interest point detection between RGB and depth images. We quantitatively show that our unsupervised method performs better or on-par with baselines.

Matching local image descriptors is a key step in many computer vision applications. For more than a decade, hand-crafted descriptors such as SIFT have been used for this task. Recently, multiple new descriptors learned from data have been proposed and shown to improve on SIFT in terms of discriminative power. This paper is dedicated to an extensive experimental evaluation of learned local features to establish a single evaluation protocol that ensures comparable results. In terms of matching performance, we evaluate the different descriptors regarding standard criteria. However, considering matching performance in isolation only provides an incomplete measure of a descriptors quality. For example, finding additional correct matches between similar images does not necessarily lead to a better performance when trying to match images under extreme viewpoint or illumination changes. Besides pure descriptor matching, we thus also evaluate the different descriptors in the context of image-based reconstruction. This enables us to study the descriptor performance on a set of more practical criteria including image retrieval, the ability to register images under strong viewpoint and illumination changes, and the accuracy and completeness of the reconstructed cameras and scenes. To facilitate future research, the full evaluation pipeline is made publicly available.

Motivated by the limitations of existing multi-view stereo benchmarks, we present a novel dataset for this task. Towards this goal, we recorded a variety of indoor and outdoor scenes using a high-precision laser scanner and captured both high-resolution DSLR imagery as well as synchronized low-resolution stereo videos with varying fields-of-view. To align the images with the laser scans, we propose a robust technique which minimizes photometric errors conditioned on the geometry. In contrast to previous datasets, our benchmark provides novel challenges and covers a diverse set of viewpoints and scene types, ranging from natural scenes to man-made indoor and outdoor environments. Furthermore, we provide data at significantly higher temporal and spatial resolution. Our benchmark is the first to cover the important use case of hand-held mobile devices while also providing high-resolution DSLR camera images. We make our datasets and an online evaluation server available at http://www.eth3d.net.

This paper deals with deep neural networks for predicting accurate dense disparity map with Semi-global matching (SGM). SGM is a widely used regularization method for real scenes because of its high accuracy and fast computation speed. Even though SGM can obtain accurate results, tuning of SGMs penalty-parameters, which control a smoothness and discontinuity of a disparity map, is uneasy and empirical methods have been proposed. We propose a learning based penalties estimation method, which we call SGM-Nets that consist of Convolutional Neural Networks. A small image patch and its position are input into SGMNets to predict the penalties for the 3D object structures. In order to train the networks, we introduce a novel loss function which is able to use sparsely annotated disparity maps such as captured by a LiDAR sensor in real environments. Moreover, we propose a novel SGM parameterization, which deploys different penalties depending on either positive or negative disparity changes in order to represent the object structures more discriminatively. Our SGM-Nets outperformed state of the art accuracy on KITTI benchmark datasets.

Consensus maximization has proven to be a useful tool for robust estimation. While randomized methods like RANSAC are fast, they do not guarantee global optimality and fail to manage large amounts of outliers. On the other hand, global methods are commonly slow because they do not exploit the structure of the problem at hand. In this paper, we show that the solution space can be reduced by introducing Linear Matrix Inequality (LMI) constraints. This leads to significant speed ups of the optimization time even for large amounts of outliers, while maintaining global optimality. We study several cases in which the objective variables have a special structure, such as rotation, scaled-rotation, and essential matrices, which are posed as LMI constraints. This is very useful in several standard computer vision problems, such as estimating Similarity Transformations, Absolute Poses, and Relative Poses, for which we obtain compelling results on both synthetic and real datasets. With up to 90 percent outlier rate, where RANSAC often fails, our constrained approach is consistently faster than the non-constrained one - while finding the same global solution.

We present a method to jointly refine the geometry and semantic segmentation of 3D surface meshes. Our method alternates between updating the shape and the semantic labels. In the geometry refinement step, the mesh is deformed with variational energy minimization, such that it simultaneously maximizes photo-consistency and the compatibility of the semantic segmentations across a set of calibrated images. Label-specific shape priors account for interactions between the geometry and the semantic labels in 3D. In the semantic segmentation step, the labels on the mesh are updated with MRF inference, such that they are compatible with the semantic segmentations in the input images. Also, this step includes prior assumptions about the surface shape of different semantic classes. The priors induce a tight coupling, where semantic information influences the shape update and vice versa. Specifically, we introduce priors that favor (i) adaptive smoothing, depending on the class label; (ii) straightness of class boundaries; and (iii) semantic labels that are consistent with the surface orientation. The novel mesh-based reconstruction is evaluated in a series of experiments with real and synthetic data. We compare both to state-of-the-art, voxel-based semantic 3D reconstruction, and to purely geometric mesh refinement, and demonstrate that the proposed scheme yields improved 3D geometry as well as an improved semantic segmentation.

Surface reconstruction from a point cloud is a standard subproblem in many algorithms for dense 3D reconstruction from RGB images or depth maps. Methods, performing only local operations in the vicinity of individual points, are very fast, but reconstructed models typically contain lots of holes. On the other hand, regularized volumetric approaches, formulated as a global optimization, are typically too slow for real-time interactive applications. We propose to use a regression forest based method, which predicts the projection of a grid point to the surface, depending on the spatial configuration of point density in the grid point neighborhood. We designed a suitable feature vector and efficient oct-tree based GPU evaluation, capable of predicting surface of high resolution 3D models in milliseconds. Our method learns and predicts surfaces from an observed point cloud sparser than the evaluation grid, and therefore effectively acts as a regularizer.

Dense depth map estimation from stereo cameras has many applications in robotic vision, e.g., obstacle detection, especially when performed in real-time. The range in which depth values can be accurately estimated is usually limited for two-camera stereo setups due to the fixed baseline between the cameras. In addition, two-camera setups suffer from wrong depth estimates caused by local minima in the matching cost functions. Both problems can be alleviated by adding more cameras as this creates multiple baselines of different lengths and since multi-image matching leads to unique minima. However, using more cameras usually comes at an increase in run-time. In this paper, we present a novel embedded system for multi-baseline stereo. By exploiting the parallelization capabilities within FPGAs, we are able to estimate a depth map from multiple cameras in real-time. We show that our approach requires only little more power and weight compared to a two-camera stereo system. At the same time, we show that our system produces significantly better depth maps and is able to handle occlusion of some cameras, resulting in the redundancy typically desired for autonomous vehicles. Our system is small in size and leight-weight and can be employed even on a MAV platform with very strict power, weight, and size requirements.

Estimating the depth from two images with a baseline has a well-known regular problem: When a line is parallel to the epipolar geometry it is not possible to estimate the depth from pixels on this line. Moreover, the classic measure for the certainty of the depth estimate fails as well: The matching score between the template and any pixel on the epipolar line is perfect. This results for common scenes in incorrect matches with very high confidence, some even resistant to left-right image checks. It is straightforward to try to address this by adding a second stereo head in a perpendicular direction. However, it is nontrivial to identify the failure and fuse the two depth maps in a real-time system. A simple weighted average will alleviate the problem but still result in a very large error in the depth map. Our contributions are: 1) We derive a model to predict the failure of stereo by leveraging the matching scores and 2) we propose a combined cost function to fuse two depth maps from orthogonal stereo heads using the failure prediction, matching score and consistency. We show the resulting system in real-time operation on a low-latency system in indoor, urban and natural environments.

Direct visual odometry and Simultaneous Localization and Mapping (SLAM) methods determine camera poses by means of direct image alignment. This optimizes a photometric cost term based on the Lucas-Kanade method. Many recent works use the brightness constancy assumption in the alignment cost formulation and therefore cannot cope with significant illumination changes. Such changes are especially likely to occur for loop closures in SLAM. Alternatives exist which attempt to match images more robustly. In our paper, we perform a systematic evaluation of real-time capable methods. We determine their accuracy and robustness in the context of odometry and of loop closures, both on real images as well as synthetic datasets with simulated lighting changes. We find that for real images, a Census-based method outperforms the others. We make our new datasets available online.

We present a direct visual odometry algorithm for a fisheye-stereo camera. Our algorithm performs simultaneous camera motion estimation and semi-dense reconstruction. The pipeline consists of two threads: a tracking thread and a mapping thread. In the tracking thread, we estimate the camera pose via semi-dense direct image alignment. To have a wider field of view (FoV) which is important for robotic perception, we use fisheye images directly without converting them to conventional pinhole images which come with a limited FoV. To address the epipolar curve problem, plane-sweeping stereo is used for stereo matching and depth initialization. Multiple depth hypotheses are tracked for selected pixels to better capture the uncertainty characteristics of stereo matching. Temporal motion stereo is then used to refine the depth and remove false positive depth hypotheses. Our implementation runs at an average of 20 Hz on a low-end PC. We run experiments in outdoor environments to validate our algorithm, and discuss the experimental results. We experimentally show that we are able to estimate 6D poses with low drift, and at the same time, do semi-dense 3D reconstruction with high accuracy. To the best of our knowledge, there is no other existing semi-dense direct visual odometry algorithm for a fisheye-stereo camera.

We present a system to generate building information models (BIMs) of house interiors from 3D scans. The strength of our approach is its simplicity and low runtime which allows for mobile processing applications. We consider scans of single floor, Manhattan-like indoor scenes for which our method creates metric room layouts by detecting walls and performing a subsequent reasoning about their neighborhood relations. The output of our method is a 3D BIM with hierarchical semantic annotations for individual rooms being refined by walls, ceilings, floors and doors. A variety of experiments demonstrate the effectiveness of our approach. Our reconstruction results compare well to other state-of-art methods in both reconstruction quality as well as runtime.

Many machine learning tasks require finding per-part correspondences between objects. In this work we focus on low-level correspondences --- a highly ambiguous matching problem. We propose to use a hierarchical semantic representation of the objects, coming from a convolutional neural network, to solve this ambiguity. Training it for low-level correspondence prediction directly might not be an option in some domains where the ground-truth correspondences are hard to obtain. We show how transfer from recognition can be used to avoid such training. Our idea is to mark parts as "matching" if their features are close to each other at all the levels of convolutional feature hierarchy (neural paths). Although the overall number of such paths is exponential in the number of layers, we propose a polynomial algorithm for aggregating all of them in a single backward pass. The empirical validation is done on the task of stereo correspondence and demonstrates that we achieve competitive results among the methods which do not use labeled target domain data.

Both image segmentation and dense 3D modeling from images represent an intrinsically ill-posed problem. Strong regularizers are therefore required to constrain the solutions from being 'too noisy'. These priors generally yield overly smooth reconstructions and/or segmentations in certain regions while they fail to constrain the solution sufficiently in other areas. In this paper, we argue that image segmentation and dense 3D reconstruction contribute valuable information to each other's task. As a consequence, we propose a mathematical framework to formulate and solve a joint segmentation and dense reconstruction problem. On the one hand knowing about the semantic class of the geometry provides information about the likelihood of the surface direction. On the other hand the surface direction provides information about the likelihood of the semantic class. Experimental results on several data sets highlight the advantages of our joint formulation. We show how weakly observed surfaces are reconstructed more faithfully compared to a geometry only reconstruction. Thanks to the volumetric nature of our formulation we also infer surfaces which cannot be directly observed for example the surface between the ground and a building. Finally, our method returns a semantic segmentation which is consistent across the whole dataset.

In this paper, we explore the different minimal solutions for egomotion estimation of a camera based on homography knowing the gravity vector between calibrated images. These solutions depend on the prior knowledge about the reference plane used by the homography. We then demonstrate that the number of matched points can vary from two to three and that a direct closed-form solution or a Gröbner basis based solution can be derived according to this plane. Many experimental results on synthetic and real sequences in indoor and outdoor environments show the efficiency and the robustness of our approach compared to standard methods.

Using synthetic aperture radar (SAR) interferometry to monitor long-term millimeter-level deformation of urban infrastructures, such as individual buildings and bridges, is an emerging and important field in remote sensing. In the state-of-the-art methods, deformation parameters are retrieved and monitored on a pixel basis solely in the SAR image domain. However, the inevitable side-looking imaging geometry of SAR results in undesired occlusion and layover in urban area, rendering the current method less competent for a semantic-level monitoring of different urban infrastructures. This paper presents a framework of a semantic-level deformation monitoring by linking the precise deformation estimates of SAR interferometry and the semantic classification labels of optical images via a 3-D geometric fusion and semantic texturing. The proposed approach provides the first “SARptical” point cloud of an urban area, which is the SAR tomography point cloud textured with attributes from optical images. This opens a new perspective of InSAR deformation monitoring. Interesting examples on bridge and railway monitoring are demonstrated.

We present a new method for decomposing an image into a set of soft color segments that are analogous to color layers with alpha channels that have been commonly utilized in modern image manipulation software. We show that the resulting decomposition serves as an effective intermediate image representation, which can be utilized for performing various, seemingly unrelated, image manipulation tasks. We identify a set of requirements that soft color segmentation methods have to fulfill, and present an in-depth theoretical analysis of prior work. We propose an energy formulation for producing compact layers of homogeneous colors and a color refinement procedure, as well as a method for automatically estimating a statistical color model from an image. This results in a novel framework for automatic and high-quality soft color segmentation that is efficient, parallelizable, and scalable. We show that our technique is superior in quality compared to previous methods through quantitative analysis as well as visually through an extensive set of examples. We demonstrate that our soft color segments can easily be exported to familiar image manipulation software packages and used to produce compelling results for numerous image manipulation applications without forcing the user to learn new tools and workflows.

We present a real-time method for rendering novel virtual camera views from given RGB-D (color and depth) data of a different viewpoint. Missing color and depth information due to incomplete input or disocclusions is efficiently inpainted in a temporally consistent way. The inpainting takes the location of strong image gradients into account as likely depth discontinuities. We present our method in the context of a view correction system for mobile devices, and discuss how to obtain a screen-camera calibration and options for acquiring depth input. Our method has use cases in both augmented and virtual reality applications. We demonstrate the speed of our system and the visual quality of its results in multiple experiments in the paper as well as in the supplementary video.

Techniques that jointly perform dense 3D reconstruction and semantic segmentation have recently shown very promising results. One major restriction so far is that they can often only handle a very low number of semantic labels. This is mostly due to their high memory consumption caused by the necessity to store indicator variables for every label and transition. We propose a way to reduce the memory consumption of existing methods. Our approach is based on the observation that many semantic labels are only present at very localized positions in the scene, such as cars. Therefore this label does not need to be active at every location. We exploit this observation by dividing the scene into blocks in which generally only a subset of labels is active. By determining early on in the reconstruction process which labels need to be active in which block the memory consumption can be significantly reduced. In order to recover from mistakes we propose to update the set of active labels during the iterative optimization procedure based on the current solution. We also propose a way to initialize the set of active labels using a boosted classifier. In our experimental evaluation we show the reduction of memory usage quantitatively. Eventually, we show results of joint semantic 3D reconstruction and semantic segmentation with significantly more labels than previous approaches were able to handle.

We present a system which is able to reconstruct human faces on mobile devices with only on-device processing using the sensors which are typically built into a current commodity smart phone. Such technology can for example be used for facial authentication purposes or as a fast preview for further post-processing. Our method uses recently proposed techniques which compute depth maps by passive multi-view stereo directly on the device. We propose an efficient method which recovers the geometry of the face from the typically noisy point cloud. First, we show that we can safely restrict the reconstruction to a 2.5D height map representation. Therefore we then propose a novel low dimensional height map shape model for faces which can be fitted to the input data efficiently even on a mobile phone. In order to be able to represent instance specific shape details, such as moles, we augment the reconstruction from the shape model with a distance map which can be regularized efficiently. We thoroughly evaluate our approach on synthetic and real data, thereby we use both high resolution depth data acquired using high quality multi-view stereo and depth data directly computed on mobile phones.

In this paper we present a deep neural network topology that incorporates a simple to implement transformationinvariant pooling operator (TI-POOLING). This operator is able to efficiently handle prior knowledge on nuisance variations in the data, such as rotation or scale changes. Most current methods usually make use of dataset augmentation to address this issue, but this requires larger number of model parameters and more training data, and results in significantly increased training time and larger chance of under-or overfitting. The main reason for these drawbacks is that that the learned model needs to capture adequate features for all the possible transformations of the input. On the other hand, we formulate features in convolutional neural networks to be transformation-invariant. We achieve that using parallel siamese architectures for the considered transformation set and applying the TI-POOLING operator on their outputs before the fully-connected layers. We show that this topology internally finds the most optimal "canonical" instance of the input image for training and therefore limits the redundancy in learned features. This more efficient use of training data results in better performance on popular benchmark datasets with smaller number of parameters when comparing to standard convolutional neural networks with dataset augmentation and to other baselines.

Capturing hyperspectral images requires expensive and specialized hardware that is not readily accessible to most users. Digital cameras, on the other hand, are significantly cheaper in comparison and can be easily purchased and used. In this paper, we present a framework for reconstructing hyperspectral images by using multiple consumer-level digital cameras. Our approach works by exploiting the different spectral sensitivities of different camera sensors. In particular, due to the differences in spectral sensitivities of the cameras, different cameras yield different RGB measurements for the same spectral signal. We introduce an algorithm that is able to combine and convert these different RGB measurements into a single hyperspectral image for both indoor and outdoor scenes. This camera-based approach allows hyperspectral imaging at a fraction of the cost of most existing hyperspectral hardware. We validate the accuracy of our reconstruction against ground truth hyperspectral images (using both synthetic and real cases) and show its usage on relighting applications.

Visual location recognition is the task of determining the place depicted in a query image from a given database of geo-tagged images. Location recognition is often cast as an image retrieval problem and recent research has almost exclusively focused on improving the chance that a relevant database image is ranked high enough after retrieval. The implicit assumption is that the number of inliers found by spatial verification can be used to distinguish between a related and an unrelated database photo with high precision. In this paper, we show that this assumption does not hold for large datasets due to the appearance of geometric bursts, i.e., sets of visual elements appearing in similar geometric configurations in unrelated database photos. We propose algorithms for detecting and handling geometric bursts. Although conceptually simple, using the proposed weighting schemes dramatically improves the recall that can be achieved when high precision is required compared to the standard re-ranking based on the inlier count. Our approach is easy to implement and can easily be integrated into existing location recognition systems.

It is well known that the rolling shutter effect in images captured with a moving rolling shutter camera causes inaccuracies to 3D reconstructions. The problem is further aggravated with weak visual connectivity from wide baseline images captured with a fast moving camera. In this paper, we propose and implement a pipeline for sparse to dense 3D construction with wide baseline images captured from a fast moving rolling shutter camera. Specifically, we propose a cost function for Bundle Adjustment (BA) that models the rolling shutter effect, incorporates GPS/INS readings, and enforces pairwise smoothness between neighboring poses. We optimize over the 3D structures, camera poses and velocities. We also introduce a novel interpolation scheme for the rolling shutter plane sweep stereo algorithm that allows us to achieve a 7× speed up in the depth map computations for dense reconstruction without losing accuracy. We evaluate our proposed pipeline over a 2.6km image sequence captured with a rolling shutter camera mounted on a moving car.

We propose an approach for dense semantic 3D reconstruction which uses a data term that is defined as potentials over viewing rays, combined with continuous surface area penalization. Our formulation is a convex relaxation which we augment with a crucial non-convex constraint that ensures exact handling of visibility. To tackle the non-convex minimization problem, we propose a majorizeminimize type strategy which converges to a critical point. We demonstrate the benefits of using the non-convex constraint experimentally. For the geometry-only case, we set a new state of the art on two datasets of the commonly used Middlebury multi-view stereo benchmark. Moreover, our general-purpose formulation directly reconstructs thin objects, which are usually treated with specialized algorithms. A qualitative evaluation on the dense semantic 3D reconstruction task shows that we improve significantly over previous methods.

This paper gives an overview of a recently proposed method of solving dense 3D reconstruction and semantic segmentation from multiple input images in a joint fashion, i.e. as semantic 3D reconstruction. The formulation is cast as a volumetric fusion of depth maps and pixel-wise semantic classification scores. By posing the two problems as a joint optimization problem, both of the tasks can benefit from the other task's information. This leads to formulations which can reconstruct hidden unobserved surfaces. We give an overview of several papers which describe different ways of modeling the data term from the input data and also works which introduce object shape priors to the formulation. We present the basic convex multi-label formulation on which the method builds and also discuss the relation to other reconstruction algorithms which extract semantically annotated 3D models from images.

The readily available image and depth data from commodity RGB-D sensors has had tremendous impact in the robotics and computer vision community recently. To jointly leverage both modalities, the depth and image measurements need to be registered. Typical calibration approaches make use of artificial landmarks and special calibration targets. However, this is not feasible if on-line (re-)calibration is necessary or the sensor setup is inaccessible, e.g., for already captured datasets. Instead of using specific calibration patterns, we propose to leverage a sparse environment model as geometric prior for the calibration. Structure-from-motion or SLAM can provide such a sparse 3D scene model, and hence our approach allows for self-calibration without the need for any manual interaction. We validate our hypothesis by introducing an optimization that jointly minimizes the alignment error between the sparse map and all recorded depth maps. Since the accuracy of depth measurements is known to degrade considerably with scene depth, we account for this distortion via a spatially varying correction term. The evaluation of our approach demonstrates that we are able to compute an accurate extrinsic and intrinsic calibration, which for example allows dense 3D modeling at improved precision.

In this paper we present Semantic Stixels, a novel vision-based scene model geared towards automated driving. Our model jointly infers the geometric and semantic layout of a scene and provides a compact yet rich abstraction of both cues using Stixels as primitive elements. Geometric information is incorporated into our model in terms of pixel-level disparity maps derived from stereo vision. For semantics, we leverage a modern deep learning-based scene labeling approach that provides an object class label for each pixel. Our experiments involve an in-depth analysis and a comprehensive assessment of the constituent parts of our approach using three public benchmark datasets. We evaluate the geometric and semantic accuracy of our model and analyze the underlying run-times and the complexity of the obtained representation. Our results indicate that the joint treatment of both cues on the Semantic Stixel level yields a highly compact environment representation while maintaining an accuracy comparable to the two individual pixel-level input data sources. Moreover, our framework compares favorably to related approaches in terms of computational costs and operates in real-time.

Automated valet parking services provide great potential to increase the attractiveness of electric vehicles by mitigating their two main current deficiencies: reduced driving ranges and prolonged refueling times. The European research project V-Charge aims at providing this service on designated parking lots using close-to-market sensors only. For this purpose the project developed a prototype capable of performing fully automated navigation in mixed traffic on designated parking lots and GPS-denied parking garages with cameras and ultrasonic sensors only. This paper summarizes the work of the project, comprising advances in network communication and parking space scheduling, multi-camera calibration, semantic mapping concepts, visual localization and motion planning. The project pushed visual localization, environment perception and automated parking to centimetre precision. The developed infrastructure-based camera calibration and semi-supervised semantic mapping concepts greatly reduce maintenance efforts. Results are presented from extensive month-long field tests.

In this paper we propose a new approach to incrementally initialize a manifold surface for automatic 3D reconstruction from images. More precisely we focus on the automatic initialization of a 3D mesh as close as possible to the final solution; indeed many approaches require a good initial solution for further refinement via multi-view stereo techniques. Our novel algorithm automatically estimates an initial manifold mesh for surface evolving multi-view stereo algorithms, where the manifold property needs to be enforced. It bootstraps from 3D points extracted via Structure from Motion, then iterates between a state-of-the-art manifold reconstruction step and a novel mesh sweeping algorithm that looks for new 3D points in the neighborhood of the reconstructed manifold to be added in the manifold reconstruction. The experimental results show quantitatively that the mesh sweeping improves the resolution and the accuracy of the manifold reconstruction, allowing a better convergence of state-of-the-art surface evolution multi-view stereo algorithms.

Due to the widespread use of compositing in contemporary feature films, green-screen keying has become an essential part of postproduction workflows. To comply with the ever-increasing quality requirements of the industry, specialized compositing artists spend countless hours using multiple commercial software tools, while eventually having to resort to manual painting because of the many shortcomings of these tools. Due to the sheer amount of manual labor involved in the process, new green-screen keying approaches that produce better keying results with less user interaction are welcome additions to the compositing artist’s arsenal. We found that—contrary to the common belief in the research community—production-quality green-screen keying is still an unresolved problem with its unique challenges. In this article, we propose a novel green-screen keying method utilizing a new energy minimization-based color unmixing algorithm. We present comprehensive comparisons with commercial software packages and relevant methods in literature, which show that the quality of our results is superior to any other currently available green-screen keying solution. It is important to note that, using the proposed method, these high-quality results can be generated using only one-tenth of the manual editing time that a professional compositing artist requires to process the same content having all previous state-of-the-art tools at one’s disposal.

This paper presents a system for 3D reconstruction of large-scale outdoor scenes based on monocular motion stereo. Ours is the first such system to run at interactive frame rates on a mobile device (Google Project Tango Tablet), thus allowing a user to reconstruct scenes "on the go" by simply walking around them. We utilize the device's GPU to compute depth maps using plane sweep stereo. We then fuse the depth maps into a global model of the environment represented as a truncated signed distance function in a spatially hashed voxel grid. We observe that in contrast to reconstructing objects in a small volume of interest, or using the near outlier-free data provided by depth sensors, one can rely less on free-space measurements for suppressing outliers in unbounded large-scale scenes. Consequently, we propose a set of simple filtering operations to remove unreliable depth estimates and experimentally demonstrate the benefit of strongly filtering depth maps. We extensively evaluate the system with real as well as synthetic datasets.

In this work we make use of recent advances in data driven classification to improve standard approaches for binocular stereo matching and single view depth estimation. Surface normal direction estimation has become feasible and shown to work reliably on state of the art benchmark datasets. Information about the surface orientation contributes crucial information about the scene geometry in cases where standard approaches struggle. We describe, how the responses of such a classifier can be included in global stereo matching approaches. One of the strengths of our approach is, that we can use the classifier responses for a whole set of directions and let the final optimization decide about the surface orientation. This is important in cases where based on the classifier, multiple different surface orientations seem likely. We evaluate our method on two challenging real-world datasets for the two proposed applications. For the binocular stereo matching we use road scene imagery taken from a car and for the single view depth estimation we use images taken in indoor environments.

Dense 3D reconstruction still remains a hard task for a broad number of object classes which are not sufficiently textured or contain transparent and reflective parts. Shape priors are the tool of choice when the input data itself is not descriptive enough to get a faithful reconstruction. We propose a novel shape prior formulation that splits the object into multiple convex parts. The reconstruction problem is posed as a volumetric multi-label segmentation. Each of the transitions between labels is penalized with its individual anisotropic smoothness term. This powerful formulation allows us to represent a descriptive shape prior. For the object classes used in this paper the individual segments naturally correspond to different semantic parts of the object. This leads to a semantic segmentation as a side product of our shape prior formulation. We evaluate our method on several challenging real-world datasets. Our results show that we can resolve issues such as undesired holes and disconnected parts. Taking into account a segmentation of the free space, we show that we are able to reconstruct concavities, such as the interior of a mug.

Videos consisting of thousands of high resolution frames are challenging for existing structure from motion (SfM) and simultaneous-localization and mapping (SLAM) techniques. We present a new approach for simultaneously computing extrinsic camera poses and 3D scene structure that is capable of handling such large volumes of image data. The key insight behind this paper is to effectively exploit coherence in densely sampled video input. Our technical contributions include robust tracking and selection of confident video frames, a novel window bundle adjustment, frame-to-structure verification for globally consistent reconstructions with multi-loop closing, and utilizing efficient global linear camera pose estimation in order to link both consecutive and distant bundle adjustment windows. To our knowledge we describe the first system that is capable of handling high resolution, high frame-rate video data with close to real-time performance. In addition, our approach can robustly integrate data from different video sequences, allowing multiple video streams to be simultaneously calibrated in an efficient and globally optimal way. We demonstrate high quality alignment on large scale challenging datasets, e.g., 2-20 megapixel resolution at frame rates of 25-120 Hz with thousands of frames.

Dense semantic 3D reconstruction is typically formulated as a discrete or continuous problem over label assignments in a voxel grid, combining semantic and depth likelihoods in a Markov Random Field framework. The depth and semantic information is incorporated as a unary potential, smoothed by a pairwise regularizer. However, modelling likelihoods as a unary potential does not model the problem correctly leading to various undesirable visibility artifacts. We propose to formulate an optimization problem that directly optimizes the reprojection error of the 3D model with respect to the image estimates, which corresponds to the optimization over rays, where the cost function depends on the semantic class and depth of the first occupied voxel along the ray. The 2-label formulation is made feasible by transforming it into a graph-representable form under QPBO relaxation, solvable using graph cut. The multi-label problem is solved by applying α-expansion using the same relaxation in each expansion move. Our method was indeed shown to be feasible in practice, running comparably fast to the competing methods, while not suffering from ray potential approximation artifacts.

We propose a novel two-step method for estimating the intrinsic and extrinsic calibration of any radially symmetric camera, including non-central systems. The first step consists of estimating the camera pose, given a Structure from Motion (SfM) model, up to the translation along the optical axis. As a second step, we obtain the calibration by finding the translation of the camera center using an ordering constraint. The method makes use of the 1D radial camera model, which allows us to effectively handle any radially symmetric camera, including non-central ones. Using this ordering constraint, we show that the we are able to calibrate several different (central and non-central) Wide Field of View (WFOV) cameras, including fisheye, hyper-catadioptric and spherical catadioptric cameras, as well as pinhole cameras, using a single image or jointly solving for several views.

Recent advances in Structure-from-Motion not only enable the reconstruction of large scale scenes, but are also able to detect ambiguous structures caused by repeating elements that might result in incorrect reconstructions. Yet, it is not always possible to fully reconstruct a scene. The images required to merge different sub-models might be missing or it might be impossible to acquire such images in the first place due to occlusions or the structure of the scene. The problem of aligning multiple reconstructions that do not have visual overlap is impossible to solve in general. An important variant of this problem is the case in which individual sides of a building can be reconstructed but not joined due to the missing visual overlap. In this paper, we present a combinatorial approach for solving this variant by automatically stitching multiple sides of a building together. Our approach exploits symmetries and semantic information to reason about the possible geometric relations between the individual models. We show that our approach is able to reconstruct complete building models where traditional SfM ends up with disconnected building sides.

Structure-based localization is the task of finding the absolute pose of a given query image w.r.t. a pre-computed 3D model. While this is almost trivial at small scale, special care must be taken as the size of the 3D model grows, because straight-forward descriptor matching becomes ineffective due to the large memory footprint of the model, as well as the strictness of the ratio test in 3D. Recently, several authors have tried to overcome these problems, either by a smart compression of the 3D model or by clever sampling strategies for geometric verification. Here we explore an orthogonal strategy, which uses all the 3D points and standard sampling, but performs feature matching implicitly, by quantization into a fine vocabulary. We show that although this matching is ambiguous and gives rise to 3D hyperpoints when matching each 2D query feature in isolation, a simple voting strategy, which enforces the fact that the selected 3D points shall be co-visible, can reliably find a locally unique 2D-3D point assignment. Experiments on two large-scale datasets demonstrate that our method achieves state-of-the-art performance, while the memory footprint is greatly reduced, since only visual word labels but no 3D point descriptors need to be stored.

Despite their enormous success in solving hard combinatorial problems, convex relaxation approaches often suffer from the fact that the computed solutions are far from binary and that subsequent heuristic binarization may substantially degrade the quality of computed solutions. In this paper, we propose a novel relaxation technique which incorporates the entropy of the objective variable as a measure of relaxation tightness. We show both theoretically and experimentally that augmenting the objective function with an entropy term gives rise to more binary solutions and consequently solutions with a substantially tighter optimality gap. We use difference of convex function (DC) programming as an efficient and provably convergent solver for the arising convex-concave minimization problem. We evaluate this approach on three prominent non-convex computer vision challenges: multi-label inpainting, image segmentation and spatio-temporal multi-view reconstruction. These experiments show that our approach consistently yields better solutions with respect to the original integral optimization problem.

The viewing graph represents a set of views that are related by pairwise relative geometries. In the context of Structure-from-Motion (SfM), the viewing graph is the input to the incremental or global estimation pipeline. Much effort has been put towards developing robust algorithms to overcome potentially inaccurate relative geometries in the viewing graph during SfM. In this paper, we take a fundamentally different approach to SfM and instead focus on improving the quality of the viewing graph before applying SfM. Our main contribution is a novel optimization that improves the quality of the relative geometries in the viewing graph by enforcing loop consistency constraints with the epipolar point transfer. We show that this optimization greatly improves the accuracy of relative poses in the viewing graph and removes the need for filtering steps or robust algorithms typically used in global SfM methods. In addition, the optimized viewing graph can be used to efficiently calibrate cameras at scale. We combine our viewing graph optimization and focal length calibration into a global SfM pipeline that is more efficient than existing approaches. To our knowledge, ours is the first global SfM pipeline capable of handling uncalibrated image sets.

Image-based localization approaches aim to determine the camera pose from which an image was taken. Finding correct 2D-3D correspondences between query image features and 3D points in the scene model becomes harder as the size of the model increases. Current state-of-the-art methods therefore combine elaborate matching schemes with camera pose estimation techniques that are able to handle large fractions of wrong matches. In this work we study the benefits and limitations of spatial verification compared to appearance-based filtering. We propose a voting-based pose estimation strategy that exhibits O(n) complexity in the number of matches and thus facilitates to consider much more matches than previous approaches - whose complexity grows at least quadratically. This new outlier rejection formulation enables us to evaluate pose estimation for 1-to-many matches and to surpass the state-of-the-art. At the same time, we show that using more matches does not automatically lead to a better performance.

This work presents a method for increasing the accuracy of standard visual inertial odometry (VIO) by effectively removing the angular drift that naturally occurs in feature-based VIO. In order to eliminate such drift, we propose to leverage the predominance of parallel lines in man-made environments by using the intersection of their image projections, known as vanishing points (VPs). First, an efficient inertial-based method is presented that accurately and efficiently detects such points. Second, a strategy to deal with these measurements within the framework of an EKF-based VIO system is presented. Furthermore, special care is taken in order to ensure the real-time execution of the estimator in order to comply with time-critical applications running on computationally constrained platforms. Experiments are performed in a mobile device on challenging environments and evaluated against the same VIO system without the use of VPs, demonstrating the superior accuracy when employing the proposed framework.

In this paper, we propose a novel and computationally efficient algorithm for simultaneous exploration and coverage with a vision-guided micro aerial vehicle (MAV) in unknown environments. This algorithm continually plans a path that allows the MAV to fulfil two objectives at the same time while avoiding obstacles: observe as much unexplored space as possible, and observe as much of the surface of the environment as possible given viewing angle and distance constraints. The former and latter objectives are known as the exploration and coverage problems respectively. Our algorithm is particularly useful for automated 3D reconstruction at the street level and in indoor environments where obstacles are omnipresent. By solving the exploration problem, we maximize the size of the reconstructed model. By solving the coverage problem, we maximize the completeness of the model. Our algorithm leverages the state lattice concept such that the planned path adheres to specified motion constraints. Furthermore, our algorithm is computationally efficient and able to run on-board the MAV in real-time. We assume that the MAV is equipped with a forward-looking depth-sensing camera in the form of either a stereo camera or RGB-D camera. We use simulation experiments to validate our algorithm. In addition, we show that our algorithm achieves a significantly higher level of coverage as compared to an exploration-only approach while still allowing the MAV to fully explore the environment.

We present a novel, deeply embedded robotics middleware and programming environment. It uses a multithreaded, publish-subscribe design pattern and provides a Unix-like software interface for micro controller applications. We improve over the state of the art in deeply embedded open source systems by providing a modular and standards-oriented platform. Our system architecture is centered around a publish-subscribe object request broker on top of a POSIX application programming interface. This allows to reuse common Unix knowledge and experience, including a bash-like shell. We demonstrate with a vertical takeoff and landing (VTOL) use case that the system modularity is well suited for novel and experimental vehicle platforms. We also show how the system architecture allows a direct interface to ROS and to run individual processes either as native ROS nodes on Linux or nodes on the micro controller, maximizing interoperability. Our microcontroller-based execution environment has substantially lower latency and better hardware connectivity than a typical Robotics Linux system and is therefore well suited for fast, high rate control tasks.

High speed, low latency obstacle avoidance is essential for enabling Micro Aerial Vehicles (MAVs) to function in cluttered and dynamic environments. While other systems exist that do high-level mapping and 3D path planning for obstacle avoidance, most of these systems require high-powered CPUs on-board or off-board control from a ground station. We present a novel entirely on-board approach, leveraging a light-weight low power stereo vision system on FPGA. Our approach runs at a frame rate of 60 frames a second on VGA-sized images and minimizes latency between image acquisition and performing reactive maneuvers, allowing MAVs to fly more safely and robustly in complex environments. We also suggest our system as a light-weight safety layer for systems undertaking more complex tasks, like mapping the environment. Finally, we show our algorithm implemented on a lightweight, very computationally constrained platform, and demonstrate obstacle avoidance in a variety of environments.

We present the design and implementation of a small Vertical-Take-Off-and-Landing (VTOL) aircraft. The vehicle requires minimal additional components to achieve the hover capability and is thus very efficient in forward flight. We improve over the state of the art by using a single controller in all flight modes without using blending between hover condition and fixed wing controllers or gain scheduling. We present a compact airflow estimation model for VTOL airframes which rely on the slipstream across control surfaces for hover attitude control. Furthermore we show attitude and position control results in simulation. Finally we show outdoor flight experiments validating our simulation results.

For autonomous navigation of Micro Aerial Vehicles (MAVs) in cluttered environments, it is essential to detect potential obstacles not only in the direction of flight but in their entire local environment. While there exist systems that do vision based obstacle detection, most of them are limited to a single perception direction. Extending these systems to a multi-directional sensing approach would exhaust the payload limit in terms of weight and computational power. We present a novel light-weight sensor setup comprising of four stereo heads and an inertial measurement unit (IMU) to perform FPGA-based dense reconstruction for obstacle detection in all directions. As the data-rate scales up with the number of cameras we use an FPGA to perform streaming based tasks in real-time and show a light-weight polar-coordinate map to allow a companion computer to fully process the data of all the cameras and perform obstacle detection in real-time. The system is able to process up to 80 frames per second (fps) freely distributed on the four stereo heads while maintaining a low power budget. The perception system including FPGA, image sensors and stereo mounts is 235 g in weight.

Mapping the environment is crucial to enable path planning and obstacle avoidance for self-driving vehicles and other robots. In this paper, we concentrate on ground-based vehicles and present an approach which extracts static obstacles from depth maps computed out of multiple consecutive images. In contrast to existing approaches, our system does not require accurate visual inertial odometry estimation but solely relies on the readily available wheel odometry. To handle the resulting higher pose uncertainty, our system fuses obstacle detections over time and between cameras to estimate the free and occupied space around the vehicle. Using monocular fisheye cameras, we are able to cover a wider field of view and detect obstacles closer to the car, which are often not within the standard field of view of a classical binocular stereo camera setup. Our quantitative analysis shows that our system is accurate enough for navigation purposes of self-driving cars and runs in real-time.

Artefacts that are present in images taken from a moving rolling shutter camera degrade the accuracy of absolute pose estimation. To alleviate this problem, we introduce an addition linear velocity in the camera projection matrix to approximate the motion of the rolling shutter camera. In particular, we derive a minimal solution using the Gröbner Basis that solves for the absolute pose as well as the motion of a rolling shutter camera. We show that the minimal problem requires 5-point correspondences and gives up to 8 real solutions. We also show that our formulation can be extended to use more than 5-point correspondences. We use RANSAC to robustly get all the inliers. In the final step, we relax the linear velocity assumption and do a non-linear refinement on the fuli motion, i.e. linear and angular velocities, and pose of the rolling shutter camera with all the inliers. We verify the feasibility and accuracy of our algorithm with both simulated and real-world datasets.

We propose a novel monocular visual inertial odometry algorithm that combines the advantages of EKF-based approaches with those of direct photometric error minimization methods. The method is based on sparse, very small patches and incorporates the minimization of photometric error directly into the EKF measurement model so that inertial data and vision-based surface measurements are used simultaneously during camera pose estimation. We fuse vision-based and inertial measurements almost at the raw-sensor level, allowing the estimated system state to constrain and guide image-space measurements. Our formulation allows for an efficient implementation that runs in real-time on a standard CPU and has several appealing and unique characteristics such as being robust to fast camera motion, in particular rotation, and not depending on the presence of corner-like features in the scene. We experimentally demonstrate robust and accurate performance compared to ground truth and show that our method works on scenes containing only non-intersecting lines.

Autonomous navigation of micro aerial vehicles (MAVs) in a-priori unknown environments is one of the most challenging problems in robotics. First, a MAV has to incrementally build a 3D geometric map from raw sensor data. Then, based on the mapping information, the path planner has to search for a cost-optimal trajectory to the goal in real-time. It is common practice to discretize the search space into a state lattice; by doing so, we reduce the path planning problem with differential constraints to a graph search problem that is easier to solve. However, a regular 3D state lattice requires a large amount of memory while graph search in a regular 3D state lattice incorporating numerous states is computationally intensive. In this paper, we introduce a novel path planning algorithm which extends the concept of a regular state lattice to an octree-based state lattice, and searches for an optimal trajectory in the octree-partitioned search space. Our octree-based state lattice representation discretizes large swathes of free space into few symbolic octants, and thus, encodes a significantly fewer number of states. As a result, memory consumption is kept to a minimum, and at the same time, graph search is made more efficient. Simulation experiments demonstrate the efficiency of path planning with an octree-based state lattice, and further field trials prove the viability of this path planning algorithm.

We propose a method to detect changes in the geometry of a city using panoramic images captured by a car driving around the city. The proposed method can be used to significantly optimize the process of updating the 3D model of an urban environment that is changing over time, by restricting this process to only those areas where changes are detected. With this application in mind, we designed our algorithm to specifically detect only structural changes in the environment, ignoring any changes in its appearance, and ignoring also all the changes which are not relevant for update purposes such as cars, people etc. The approach also accounts for the challenges involved in a large scale application of change detection, such as inaccuracies in the input geometry, errors in the geo-location data of the images as well as the limited amount of information due to sparse imagery. We evaluated our approach on a small scale setup using high resolution, densely captured images and a large scale setup covering an entire city using instead the more realistic scenario of low resolution, sparsely captured images. A quantitative evaluation was also conducted for the large scale setup consisting of 14,000 images.

Traditional fluid simulations require large computational resources even for an average sized scene with the main bottleneck being a very small time step size, required to guarantee the stability of the solution. Despite a large progress in parallel computing and efficient algorithms for pressure computation in the recent years, realtime fluid simulations have been possible only under very restricted conditions. In this paper we propose a novel machine learning based approach, that formulates physics-based fluid simulation as a regression problem, estimating the acceleration of every particle for each frame. We designed a feature vector, directly modelling individual forces and constraints from the Navier-Stokes equations, giving the method strong generalization properties to reliably predict positions and velocities of particles in a large time step setting on yet unseen test videos. We used a regression forest to approximate the behaviour of particles observed in the large training set of simulations obtained using a traditional solver. Our GPU implementation led to a speed-up of one to three orders of magnitude compared to the state-of-the-art position-based fluid solver and runs in real-time for systems with up to 2 million particles.

We propose a new approach to estimate the geometric extrinsic calibration of all the elements of a smart phone or tablet (such as the screen, the front and the back cameras) by using a planar mirror. By moving a smart phone in front of a single static planar mirror, it is possible to establish correspondences between the images and a pattern displayed on the screen, and therefore estimate the geometric relationship between the non-overlapping cameras with respect to the screen location. The newly proposed setup (static mirror, moving smart phone) enables to both improve the state-of-the-art by working in the minimal case of two images, and improve the accuracy when more images are available. We analyze the minimal case for different calibration scenarios and evaluate the proposed approach on several data. We also show an application of this geometric calibration for specular surface reconstruction, by observing the reflection of a known pattern displayed on the screen.

In this paper, we propose an adaptation of camera projection models for fisheye cameras into the plane-sweeping stereo matching algorithm. This adaptation allows us to do plane-sweeping stereo directly on fisheye images. Our approach also works for other non-pinhole cameras such as omni directional and catadioptric cameras when using the unified projection model. Despite the simplicity of our proposed approach, we are able to obtain full, good quality and high resolution depth maps from the fisheye images. To verify our approach, we show experimental results based on depth maps generated by our approach, and dense models produced from these depth maps.

A novel method for auto-calibration of a PTZ (pan-tilt-zoom) camera network is proposed. The key idea on which it is based is to use pan-tilt motions generated by PTZ cameras themselves as calibration patterns. Generating and observing the pan-tilt motions of each camera makes it possible to estimate whole relative camera poses in a camera network. Cameras first observe circular trajectories of a marker set on another camera performing tilt motions with various pan angles. Although finding correspondences between marker points captured by the cameras is cumbersome due to network delays, that of between circular trajectories is easy. It is thus possible to use several hundred points in a circular trajectory to estimate its normal vectors and centers with high accuracy. Imposing geometric constraints on them makes it possible to eliminate ambiguities of them and obtain a unique relative camera pose directly. A novel refinement process to minimize differences between modeled pan-tilt motions and three or more circular trajectories is then performed. Experiments using synthetic data confirmed that the refinement process improved camera pose estimation accuracy in comparison to the direct estimation. An experiment using real data showed that the proposed method works properly for real cameras.

Key point matching in images of indoor scenes traditionally employs features like SIFT, GIST and HOG. While those features work very well for two images related to each other by small camera transformations, we commonly observe a drop in performance for patches representing scene elements visualized from a very different perspective. Since increasing the space of considered local transformations for feature matching decreases their discriminative abilities, we propose a more global approach inspired by the recent success of monocular scene understanding. In particular we propose to reconstruct a box-like model of the scene from every single image and use it to rectify images before matching. We show that a monocular scene model reconstruction and rectification preceding standard feature matching significantly improves key point matching and dramatically improves reconstruction of difficult indoor scenes.

In this work we propose an optimization scheme for variational, vectorial denoising and fusion of surface normal maps. These are common outputs of shape from shading, photometric stereo or single image reconstruction methods, but tend to be noisy and request post-processing for further usage. Processing of normals maps, which do not provide knowledge about the underlying scene depth, is complicated due to their unit length constraint which renders the optimization non-linear and non-convex. The presented approach builds upon a linearization of the constraint to obtain a convex relaxation, while guaranteeing convergence. Experimental results demonstrate that our algorithm generates more consistent representations from estimated and potentially complementary normal maps.

We propose a sequential optimization technique for segmenting a rectified image of a facade into semantic categories. Our method retrieves a parsing which respects common architectural constraints and also returns a certificate for global optimality. Contrasting the suggested method, the considered facade labeling problem is typically tackled as a classification task or as grammar parsing. Both approaches are not capable of fully exploiting the regularity of the problem. Therefore, our technique very significantly improves the accuracy compared to the state-of-the-art while being an order of magnitude faster. In addition, in 85% of the test images we obtain a certificate for optimality.

Motivated by a Bayesian vision of the 3D multi-view reconstruction from images problem, we propose a dense 3D reconstruction technique that jointly refines the shape and the camera parameters of a scene by minimizing the photometric reprojection error between a generated model and the observed images, hence considering all pixels in the original images. The minimization is performed using a gradient descent scheme coherent with the shape representation (here a triangular mesh), where we derive evolution equations in order to optimize both the shape and the camera parameters. This can be used at a last refinement step in 3D reconstruction pipelines and helps improving the 3D reconstruction's quality by estimating the 3D shape and camera calibration more accurately. Examples are shown for multi-view stereo where the texture is also jointly optimized and improved, but could be used for any generative approaches dealing with multi-view reconstruction settings (ie depth map fusion, multi-view photometric stereo).

Dense 3D reconstruction of real world objects containing textureless, reflective and specular parts is a challenging task. Using general smoothness priors such as surface area regularization can lead to defects in the form of disconnected parts or unwanted indentations. We argue that this problem can be solved by exploiting the object class specific local surface orientations, e.g. a car is always close to horizontal in the roof area. Therefore, we formulate an object class specific shape prior in the form of spatially varying anisotropic smoothness terms. The parameters of the shape prior are extracted from training data. We detail how our shape prior formulation directly fits into recently proposed volumetric multi-label reconstruction approaches. This allows a segmentation between the object and its supporting ground. In our experimental evaluation we show reconstructions using our trained shape prior on several challenging datasets.

In this paper, we propose an efficient and accurate scheme for the integration of multiple stereo-based depth measurements. For each provided depth map a confidence-based weight is assigned to each depth estimate by evaluating local geometry orientation, underlying camera setting and photometric evidence. Subsequently, all hypotheses are fused together into a compact and consistent 3D model. Thereby, visibility conflicts are identified and resolved, and fitting measurements are averaged with regard to their confidence scores. The individual stages of the proposed approach are validated by comparing it to two alternative techniques which rely on a conceptually different fusion scheme and a different confidence inference, respectively. Pursuing live 3D reconstruction on mobile devices as a primary goal, we demonstrate that the developed method can easily be integrated into a system for monocular interactive 3D modeling by substantially improving its accuracy while adding a negligible overhead to its performance and retaining its interactive potential.

The limitations of current state-of-the-art methods for single-view depth estimation and semantic segmentations are closely tied to the property of perspective geometry, that the perceived size of the objects scales inversely with the distance. In this paper, we show that we can use this property to reduce the learning of a pixel-wise depth classifier to a much simpler classifier predicting only the likelihood of a pixel being at an arbitrarily fixed canonical depth. The likelihoods for any other depths can be obtained by applying the same classifier after appropriate image manipulations. Such transformation of the problem to the canonical depth removes the training data bias towards certain depths and the effect of perspective. The approach can be straight-forwardly generalized to multiple semantic classes, improving both depth estimation and semantic segmentation performance by directly targeting the weaknesses of independent approaches. Conditioning the semantic label on the depth provides a way to align the data to their physical scale, allowing to learn a more discriminative classifier. Conditioning depth on the semantic class helps the classifier to distinguish between ambiguities of the otherwise ill-posed problem. We tested our algorithm on the KITTI road scene dataset and NYU2 indoor dataset and obtained obtained results that significantly outperform current state-of-the-art in both single-view depth and semantic segmentation domain.

In this paper, we present our minimal 4-point and linear 8-point algorithms to estimate the relative pose of a multi-camera system with known vertical directions, i.e. known absolute roll and pitch angles. We solve the minimal 4-point algorithm with the hidden variable resultant method and show that it leads to an 8-degree univariate polynomial that gives up to 8 real solutions. We identify a degenerated case from the linear 8-point algorithm when it is solved with the standard Singular Value Decomposition (SVD) method and adopt a simple alternative solution which is easy to implement. We show that our proposed algorithms can be efficiently used within RANSAC for robust estimation. We evaluate the accuracy of our proposed algorithms by comparisons with various existing algorithms for the multi-camera system on simulations and show the feasibility of our proposed algorithms with results from multiple real-world datasets.

While MAP inference is typically intractable for many real-world applications, linear programming relaxations have been proven very effective. Dual block-coordinate descent methods are among the most efficient solvers, however, they are prone to get stuck in sub-optimal points. Although subgradient approaches achieve global convergence, they are typically slower in practice. To improve convergence speed, algorithms which compute the steepest ε-descent direction by solving a quadratic program have been proposed. In this paper we suggest to decouple the quadratic program based on the Frank-Wolfe approach. This allows us to obtain an efficient and easy to parallelize algorithm while retaining the global convergence properties. Our method proves superior when compared to existing algorithms on a set of spin-glass models and protein design tasks.

The online recalibration of multi-sensor systems is a fundamental problem that must be solved before complex automated systems are deployed in situations such as automated driving. In such situations, accurate knowledge of calibration parameters is critical for the safe operation of automated systems. However, most existing calibration methods for multisensor systems are computationally expensive, use installations of known fiducial patterns, and require expert supervision. We propose an alternative approach called infrastructure-based calibration that is efficient, requires no modification of the infrastructure, and is completely unsupervised. In a survey phase, a computationally expensive simultaneous localization and mapping (SLAM) method is used to build a highly accurate map of a calibration area. Once the map is built, many other vehicles are able to use it for calibration as if it were a known fiducial pattern. We demonstrate the effectiveness of this method to calibrate the extrinsic parameters of a multi-camera system. The method does not assume that the cameras have an overlapping field of view and it does not require an initial guess. As the camera rig moves through the previously mapped area, we match features between each set of synchronized camera images and the map. Subsequently, we find the camera poses and inlier 2D-3D correspondences. From the camera poses, we obtain an initial estimate of the camera extrinsics and rig poses, and optimize these extrinsics and rig poses via non-linear refinement. The calibration code is publicly available as a standalone C++ package.

A potential loop-closure image pair passes the geometric verification test if the number of inliers from the computation of the geometric constraint with RANSAC exceed a pre-defined threshold. The choice of the threshold is critical to the success of identifying the correct loop-closure image pairs. However, the value for this threshold often varies for different datasets and is chosen empirically. In this paper, we propose an unsupervised method that learns the threshold for geometric verification directly from the observed inlier counts of all the potential loop-closure image pairs. We model the distributions of the inlier counts from all the potential loop-closure image pairs with a two components Log-Normal mixture model - one component represents the state of non loop-closure and the other represents the state of loop-closure, and learn the parameters with the Expectation-Maximization algorithm. The intersection of the Log-Normal mixture distributions is the optimal threshold for geometric verification, i.e. the threshold that gives the minimum false positive and negative loop-closures. Our algorithm degenerates when there are too few or no loop-closures and we propose the χ2test to detect this degeneracy. We verify our proposed method with several large-scale datasets collected from both the multi-camera setup and stereo camera.

Recent developments in smartphones create an ideal platform for robotics and computer vision applications: they are small, powerful, embedded devices with low-power mobile CPUs. However, though the computational power of smartphones has increased substantially in recent years, they are still not capable of performing intense computer vision tasks in real time, at high frame rates and low latency. We present a combination of FPGA and mobile CPU to overcome the computational and latency limitations of mobile CPUs alone. With the FPGA as an additional layer between the image sensor and CPU, the system is capable of accelerating computer vision algorithms to real-time performance. Low latency calculation allows for direct usage within control loops of mobile robots. A stereo camera setup with disparity estimation based on the semi global matching algorithm is implemented as an accelerated example application. The system calculates dense disparity images with 752×480 pixels resolution at 60 frames per second. The overall latency of the disparity estimation is less than 2 milliseconds. The system is suitable for any mobile robot application due to its light weight and low power consumption.

In this work, we present a unified view on Markov random fields (MRFs) and recently proposed continuous tight convex relaxations for multilabel assignment in the image plane. These relaxations are far less biased toward the grid geometry than Markov random fields on grids. It turns out that the continuous methods are nonlinear extensions of the well-established local polytope MRF relaxation. In view of this result, a better understanding of these tight convex relaxations in the discrete setting is obtained. Further, a wider range of optimization methods is now applicable to find a minimizer of the tight formulation. We propose two methods to improve the efficiency of minimization. One uses a weaker, but more efficient continuously inspired approach as initialization and gradually refines the energy where it is necessary. The other one reformulates the dual energy enabling smooth approximations to be used for efficient optimization. We demonstrate the utility of our proposed minimization schemes in numerical experiments. Finally, we generalize the underlying energy formulation from isotropic metric smoothness costs to arbitrary nonmetric and orientation dependent smoothness terms.

Given the growth of Internet photo collections, we now have a visual index of all major cities and tourist sites in the world. However, it is still a difficult task to capture that perfect shot with your own camera when visiting these places, especially when your camera itself has limitations, such as a limited field of view. In this paper, we propose a framework to overcome the imperfections of personal photographs of tourist sites using the rich information provided by large-scale Internet photo collections. Our method deploys state-of-the-art techniques for constructing initial 3D models from photo collections. The same techniques are then used to register personal photographs to these models, allowing us to augment personal 2D images with 3D information. This strong available scene prior allows us to address a number of traditionally challenging image enhancement techniques and achieve high-quality results using simple and robust algorithms. Specifically, we demonstrate automatic foreground segmentation, mono-to-stereo conversion, field-of-view expansion, photometric enhancement, and additionally automatic annotation with geolocation and tags. Our method clearly demonstrates some possible benefits of employing the rich information contained in online photo databases to efficiently enhance and augment one's own personal photographs.

We present a near real-time algorithm for interactively exploring a collectively captured moment without explicit 3D reconstruction. Our system favors immediacy and local coherency to global consistency. It is common to represent photos as vertices of a weighted graph, where edge weights measure similarity or distance between pairs of photos. We introduce Angled Graphs as a new data structure to organize collections of photos in a way that enables the construction of visually smooth paths. Weighted angled graphs extend weighted graphs with angles and angle weights which penalize turning along paths. As a result, locally straight paths can be computed by specifying a photo and a direction. The weighted angled graphs of photos used in this paper can be regarded as the result of discretizing the Riemannian geometry of the high dimensional manifold of all possible photos. Ultimately, our system enables everyday people to take advantage of each others' perspectives in order to create on-the-spot spatiotemporal visual experiences similar to the popular bullet-time sequence. We believe that this type of application will greatly enhance shared human experiences spanning from events as personal as parents watching their children's football game to highly publicized red carpet galas.

This paper presents a computational effective framework for tracking and pose estimation of vehicles in videos reaching comparable performance to state-of-the-art methods. We cast the problem of vehicle tracking as ranking possible poses for each frame and connecting subsequent poses by exploiting a feasible motion model over time. As a novelty, we use random forests trained on a set of existing 3D models for estimating the pose. We discretize the viewpoint space for training, where a synthetic camera is orbiting around the models. To compare projections of 3D models to real world 2D input frames, we introduce simple but discriminative principle gradient features to describe both images. A Markov Random Field ensures to pick the perfect pose over time and the vehicle to follow a feasible motion. As can be seen from our experiments performed on a variety of videos with vast variation of vehicle types, the proposed framework achieves similar results in less computational time compared to state-of-the-art methods.

In cameras with radial distortion, straight lines in space are in general mapped to curves in the image. Although epipolar geometry also gets distorted, there is a set of special epipolar lines that remain straight, namely those that go through the distortion center. By finding these straight epipolar lines in camera pairs we can obtain constraints on the distortion center(s) without any calibration object or plumb line assumptions in the scene. Although this holds for all radial distortion models we conceptually prove this idea using the division distortion model and the radial fundamental matrix which allow for a very simple closed form solution of the distortion center from two views (same distortion) or three views (different distortions). The non-iterative nature of our approach makes it immune to local minima and allows finding the distortion center also for cropped images or those where no good prior exists. Besides this, we give comprehensive relations between different undistortion models and discuss advantages and drawbacks.

Both image segmentation and dense 3D modeling from images represent an intrinsically ill-posed problem. Strong regularizers are therefore required to constrain the solutions from being 'too noisy'. Unfortunately, these priors generally yield overly smooth reconstructions and/or segmentations in certain regions whereas they fail in other areas to constrain the solution sufficiently. In this paper we argue that image segmentation and dense 3D reconstruction contribute valuable information to each other's task. As a consequence, we propose a rigorous mathematical framework to formulate and solve a joint segmentation and dense reconstruction problem. Image segmentations provide geometric cues about which surface orientations are more likely to appear at a certain location in space whereas a dense 3D reconstruction yields a suitable regularization for the segmentation problem by lifting the labeling from 2D images to 3D space. We show how appearance-based cues and 3D surface orientation priors can be learned from training data and subsequently used for class-specific regularization. Experimental results on several real data sets highlight the advantages of our joint formulation.

Articulated objects represent an important class of objects in our everyday environment. Automatic detection of the type of articulated or otherwise restricted motion and extraction of the corresponding motion parameters are therefore of high value, \eg in order to augment an otherwise static 3D reconstruction with dynamic semantics, such as rotation axes and allowable translation directions for certain rigid parts or objects. Hence, in this paper, a novel theory to analyse relative transformations between two motion-restricted parts will be presented. The analysis is based on linear subspaces spanned by relative transformations. Moreover, a signature for relative transformations will be introduced which uniquely specifies the type of restricted motion encoded in these relative transformations. This theoretic framework enables the derivation of novel algebraic constraints, such as low-rank constraints for subsequent rotations around two fixed axes for example. Lastly, given the type of restricted motion as predicted by the signature, the paper shows how to extract all the motion parameters with matrix manipulations from linear algebra. Our theory is verified on several real data sets, such as a rotating blackboard or a wheel rolling on the floor amongst others.

In this paper, we present a visual ego-motion estimation algorithm for a self-driving car equipped with a close-to-market multi-camera system. By modeling the multi-camera system as a generalized camera and applying the non-holonomic motion constraint of a car, we show that this leads to a novel 2-point minimal solution for the generalized essential matrix where the full relative motion including metric scale can be obtained. We provide the analytical solutions for the general case with at least one inter-camera correspondence and a special case with only intra-camera correspondences. We show that up to a maximum of 6 solutions exist for both cases. We identify the existence of degeneracy when the car undergoes straight motion in the special case with only intra-camera correspondences where the scale becomes unobservable and provide a practical alternative solution. Our formulation can be efficiently implemented within RANSAC for robust estimation. We verify the validity of our assumptions on the motion model by comparing our results on a large real-world dataset collected by a car equipped with 4 cameras with minimal overlapping field-of-views against the GPS/INS ground truth.

In this paper, we propose a method to detect changes in the geometry of a city using panoramic images captured by a car driving around the city. We designed our approach to account for all the challenges involved in a large scale application of change detection, such as, inaccuracies in the input geometry, errors in the geo-location data of the images, as well as, the limited amount of information due to sparse imagery. We evaluated our approach on an area of 6 square kilometers inside a city, using 3420 images downloaded from Google Street View. These images besides being publicly available, are also a good example of panoramic images captured with a driving vehicle, and hence demonstrating all the possible challenges resulting from such an acquisition. We also quantitatively compared the performance of our approach with respect to a ground truth, as well as to prior work. This evaluation shows that our approach outperforms the current state of the art.

Although specular objects have gained interest in recent years, virtually no approaches exist for marker less reconstruction of reflective scenes in the wild. In this work, we present a practical approach to capturing normal maps in real-world scenes using video only. We focus on nearly planar surfaces such as windows, facades from glass or metal, or frames, screens and other indoor objects and show how normal maps of these can be obtained without the use of an artificial calibration object. Rather, we track the reflections of real-world straight lines, while moving with a hand-held or vehicle-mounted camera in front of the object. In contrast to error-prone local edge tracking, we obtain the reflections by a robust, global segmentation technique of an ortho-rectified 3D video cube that also naturally allows efficient user interaction. Then, at each point of the reflective surface, the resulting 2D-curve to 3D-line correspondence provides a novel quadratic constraint on the local surface normal. This allows to globally solve for the shape by integrability and smoothness constraints and easily supports the usage of multiple lines. We demonstrate the technique on several objects and facades.

A huge fraction of cameras used nowadays is based on CMOS sensors with a rolling shutter that exposes the image line by line. For dynamic scenes/cameras this introduces undesired effects like stretch, shear and wobble. It has been shown earlier that rotational shake induced rolling shutter effects in hand-held cell phone capture can be compensated based on an estimate of the camera rotation. In contrast, we analyse the case of significant camera motion, e.g.\ where a bypassing street level capture vehicle uses a rolling shutter camera in a 3D reconstruction framework. The introduced error is depth dependent and cannot be compensated based on camera motion/rotation alone, invalidating also rectification for stereo camera systems. On top, significant lens distortion as often present in wide angle cameras intertwines with rolling shutter effects as it changes the time at which a certain 3D point is seen. We show that naive 3D reconstructions (assuming global shutter) will deliver biased geometry already for very mild assumptions on vehicle speed and resolution. We then develop rolling shutter dense multiview stereo algorithms that solve for time of exposure and depth at the same time, even in the presence of lens distortion and perform an evaluation on ground truth laser scan models as well as on real street-level data.

In this paper we propose an approach to jointly infer the room layout as well as the objects present in the scene. Towards this goal, we propose a branch and bound algorithm which is guaranteed to retrieve the global optimum of the joint problem. The main difficulty resides in taking into account occlusion in order to not over-count the evidence. We introduce a new decomposition method, which generalizes integral geometry to triangular shapes, and allows us to bound the different terms in constant time. We exploit both geometric cues and object detectors as image features and show large improvements in 2D and 3D object detection over state-of-the-art deformable part-based models.

In this paper, we propose a complete on-device 3D reconstruction pipeline for mobile monocular hand-held devices, which generates dense 3D models with absolute scale on-site while simultaneously supplying the user with real-time interactive feedback. The method fills a gap in current cloud-based mobile reconstruction services as it ensures at capture time that the acquired image set fulfills desired quality and completeness criteria. In contrast to existing systems, the developed framework offers multiple innovative solutions. In particular, we investigate the usability of the available on-device inertial sensors to make the tracking and mapping process more resilient to rapid motions and to estimate the metric scale of the captured scene. Moreover, we propose an efficient and accurate scheme for dense stereo matching which allows to reduce the processing time to interactive speed. We demonstrate the performance of the reconstruction pipeline on multiple challenging indoor and outdoor scenes of different size and depth variability.

We address the problem of wide-baseline registration of RGB-D data, such as photo-textured laser scans without any artificial targets or prediction on the relative motion. Our approach allows to fully automatically register scans taken in GPS-denied environments such as urban canyon, industrial facilities or even indoors. We build upon image features which are plenty, localized well and much more discriminative than geometry features, however, they suffer from viewpoint distortions and request for normalization. We utilize the principle of salient directions present in the geometry and propose to extract (several) directions from the distribution of surface normals or other cues such as observable symmetries. Compared to previous work we pose no requirements on the scanned scene (like containing large textured planes) and can handle arbitrary surface shapes. Rendering the whole scene from these repeatable directions using an orthographic camera generates textures which are identical up to 2D similarity transformations. This ambiguity is naturally handled by 2D features and allows to find stable correspondences among scans. For geometric pose estimation from tentative matches we propose a fast and robust 2 point sample consensus scheme integrating an early rejection phase. We evaluate our approach on different challenging real world scenes.

Robust velocity and position estimation at high update rates is crucial for mobile robot navigation. In recent years optical flow sensors based on computer mouse hardware chips have been shown to perform well on micro air vehicles. Since they require more light than present in typical indoor and outdoor low-light conditions, their practical use is limited. We present an open source and open hardware design 1 of an optical flow sensor based on a machine vision CMOS image sensor for indoor and outdoor applications with very high light sensitivity. Optical flow is estimated on an ARM Cortex M4 microcontroller in real-time at 250 Hz update rate. Angular rate compensation with a gyroscope and distance scaling using a ultrasonic sensor are performed onboard. The system is designed for further extension and adaption and shown in-flight on a micro air vehicle.

Multiple cameras are increasingly prevalent on robotic and human-driven vehicles. These cameras come in a variety of wide-angle, fish-eye, and catadioptric models. Furthermore, wheel odometry is generally available on the vehicles on which the cameras are mounted. For robustness, vision applications tend to use wheel odometry as a strong prior for camera pose estimation, and in these cases, an accurate extrinsic calibration is required in addition to an accurate intrinsic calibration. To date, there is no known work on automatic intrinsic calibration of generic cameras, and more importantly, automatic extrinsic calibration of a rig with multiple generic cameras and odometry. We propose an easy-to-use automated pipeline that handles both intrinsic and extrinsic calibration; we do not assume that there are overlapping fields of view. At the begining, we run an intrinsic calibration for each generic camera. The intrinsic calibration is automatic and requires a chessboard. Subsequently, we run an extrinsic calibration which finds all camera-odometry transforms. The extrinsic calibration is unsupervised, uses natural features, and only requires the vehicle to be driven around for a short time. The intrinsic parameters are optimized in a final bundle adjustment step in the extrinsic calibration. In addition, the pipeline produces a globally-consistent sparse map of landmarks which can be used for visual localization. The pipeline is publicly available as a standalone C++ package.

In this paper, we model the robust loop-closure pose-graph SLAM problem as a Bayesian network and show that it can be solved with the Classification Expectation-Maximization (EM) algorithm. In particular, we express our robust pose-graph SLAM as a Bayesian network where the robot poses and constraints are latent and observed variables. An additional set of latent variables is introduced as weights for the loop-constraints. We show that the weights can be chosen as the Cauchy function, which are iteratively computed from the errors between the predicted robot poses and observed loop-closure constraints in the Expectation step, and used to weigh the cost functions from the pose-graph loop-closure constraints in the Maximization step. As a result, outlier loop-closure constraints are assigned low weights and exert less influences in the pose-graph optimization within the EM iterations. To prevent the EM algorithm from getting stuck at local minima, we perform the EM algorithm multiple times where the loop constraints with very low weights are removed after each EM process. This is repeated until there are no more changes to the weights. We show proofs of the conceptual similarity between our EM algorithm and the M-Estimator. Specifically, we show that the weight function in our EM algorithm is equivalent to the robust residual function in the M-Estimator. We verify our proposed algorithm with experimental results from multiple simulated and real-world datasets, and comparisons with other existing works.

In this paper, we propose a method to compute the pose-graph loop-closure constraints using multiple non/minimal overlapping field-of-views cameras mounted rigidly on a self-driving car without the need to reconstruct any 3D scene points. In particular, we show that the relative pose with metric scale between two loop-closing pose-graph vertices can be directly obtained from the epipolar geometry of the multicameras system. As a result, we avoid the additional time complexities and uncertainties from the reconstruction of 3D scene points which are needed by standard monocular and stereo approaches. In addition, there is a greater flexibility in choosing a configuration for the multi-camera system to cover a wider field-of-view so as to avoid missing out any loop-closure opportunities. We show that by expressing the point correspondences between two frames as Plücker lines and enforcing the planar motion constraint on the car, we are able to use multiple cameras as one and formulate the relative pose problem for loop-closure as a minimal problem which requires 3-point correspondences that yields up to six real solutions. The RANSAC algorithm is used to determine the correct solution and for robust estimation. We verify our method with results from multiple large-scale real-world data.

This paper presents a novel feature descriptor-based calibration pattern and a Matlab toolbox which uses the specially designed pattern to easily calibrate both the intrin-sics and extrinsics of a multiple-camera system. In contrast to existing calibration patterns, in particular, the ubiquitous chessboard, the proposed pattern contains many more features of varying scales; such features can be easily and automatically detected. The proposed toolbox supports the calibration of a camera system which can comprise either normal pinhole cameras or catadioptric cameras. The calibration only requires that neighboring cameras observe parts of the calibration pattern at the same time; the observed parts may not overlap at all. No overlapping fields of view are assumed for the camera system. We show that the toolbox can easily be used to automatically calibrate camera systems.

We propose an algorithm to estimate the relative camera pose using four feature correspondences and one relative rotation angle measurement. The algorithm can be used for relative pose estimation of a rigid body equipped with a camera and a relative rotation angle sensor which can be either an odometer, an IMU or a GPS/INS system. This algorithm exploits the fact that the relative rotation angles of both the camera and relative rotation angle sensor are the same as the camera and sensor are rigidly mounted to a rigid body. Therefore, knowledge of the extrinsic calibration between the camera and sensor is not required. We carry out a quantitative comparison of our algorithm with the well-known 5-point and 1-point algorithms, and show that our algorithm exhibits the highest level of accuracy.

Future requirements for drastic reduction of CO2production and energy consumption will lead to significant changes in the way we see mobility in the years to come. However, the automotive industry has identified significant barriers to the adoption of electric vehicles, including reduced driving range and greatly increased refueling times. Automated cars have the potential to reduce the environmental impact of driving, and increase the safety of motor vehicle travel. The current state-of-the-art in vehicle automation requires a suite of expensive sensors. While the cost of these sensors is decreasing, integrating them into electric cars will increase the price and represent another barrier to adoption. The V-Charge Project, funded by the European Commission, seeks to address these problems simultaneously by developing an electric automated car, outfitted with close-to-market sensors, which is able to automate valet parking and recharging for integration into a future transportation system. The final goal is the demonstration of a fully operational system including automated navigation and parking. This paper presents an overview of the V-Charge system, from the platform setup to the mapping, perception, and planning sub-systems.

We present a supervised learning-based method to estimate a per-pixel confidence for optical flow vectors. Regions of low texture and pixels close to occlusion boundaries are known to be difficult for optical flow algorithms. Using a spatiotemporal feature vector, we estimate if a flow algorithm is likely to fail in a given region. Our method is not restricted to any specific class of flow algorithm and does not make any scene specific assumptions. By automatically learning this confidence, we can combine the output of several computed flow fields from different algorithms to select the best performing algorithm per pixel. Our optical flow confidence measure allows one to achieve better overall results by discarding the most troublesome pixels. We illustrate the effectiveness of our method on four different optical flow algorithms over a variety of real and synthetic sequences. For algorithm selection, we achieve the top overall results on a large test set, and at times even surpass the results of the best algorithm among the candidates.

A computational problem that arises frequently in computer vision is that of estimating the parameters of a model from data that have been contaminated by noise and outliers. More generally, any practical system that seeks to estimate quantities from noisy data measurements must have at its core some means of dealing with data contamination. The random sample consensus (RANSAC) algorithm is one of the most popular tools for robust estimation. Recent years have seen an explosion of activity in this area, leading to the development of a number of techniques that improve upon the efficiency and robustness of the basic RANSAC algorithm. In this paper, we present a comprehensive overview of recent research in RANSAC-based robust estimation by analyzing and comparing various approaches that have been explored over the years. We provide a common context for this analysis by introducing a new framework for robust estimation, which we call Universal RANSAC (USAC). USAC extends the simple hypothesize-and-verify structure of standard RANSAC to incorporate a number of important practical and computational considerations. In addition, we provide a general-purpose C++ software library that implements the USAC framework by leveraging state-of-the-art algorithms for the various modules. This implementation thus addresses many of the limitations of standard RANSAC within a single unified package. We benchmark the performance of the algorithm on a large collection of estimation problems. The implementation we provide can be used by researchers either as a stand-alone tool for robust estimation or as a benchmark for evaluating new techniques.

For modern consumer cameras often approximate calibration data is available, mak-ing applications such as 3D reconstruction or photo registration easier as compared to the pure uncalibrated setting. In this paper we address the setting with calibrated-uncalibrated image pairs: for one image intrinsic parameters are assumed to be known, whereas the second view has unknown distortion and calibration parameters. This sit-uation arises e.g. when one would like to register archive imagery to recently taken photos. A commonly adopted strategy for determining epipolar geometry is based on feature matching and minimal solvers inside a RANSAC framework. However, only very few existing solutions apply to the calibrated-uncalibrated setting. We propose a simple and numerically stable two-step scheme to first estimate radial distortion parameters and subsequently the focal length using novel solvers. We demonstrate the performance on synthetic and real datasets. 1

The projections of world parallel lines in an image intersect at a single point called the vanishing point (VP). VPs are a key ingredient for various vision tasks including rotation estimation and 3D reconstruction. Urban environments generally exhibit some dominant orthogonal VPs. Given a set of lines extracted from a calibrated image, this paper aims to (1) determine the line clustering, i.e. find which line belongs to which VP, and (2) estimate the associated orthogonal VPs. None of the existing methods is fully satisfactory because of the inherent difficulties of the problem, such as the local minima and the chicken-and-egg aspect. In this paper, we present a new algorithm that solves the problem in a mathematically guaranteed globally optimal manner and can inherently enforce the VP orthogonality. Specifically, we formulate the task as a consensus set maximization problem over the rotation search space, and further solve it efficiently by a branch-and-bound procedure based on the Interval Analysis theory. Our algorithm has been validated successfully on sets of challenging real images as well as synthetic data sets.

Many architectural scenes contain symmetric or repeated structures, which can generate erroneous image correspondences during structure from motion (Sfm) computation. Prior work has shown that the detection and removal of these incorrect matches is crucial for accurate and robust recovery of scene structure. In this paper, we point out that these incorrect matches, in fact, provide strong cues to the existence of symmetries and structural regularities in the unknown 3D structure. We make two key contributions. First, we propose a method to recover various symmetry relations in the structure using geometric and appearance cues. A set of structural constraints derived from the symmetries are imposed within a new constrained bundle adjustment formulation, where symmetry priors are also incorporated. Second, we show that the recovered symmetries enable us to choose a natural coordinate system for the 3D structure where gauge freedom in rotation is held fixed. Furthermore, based on the symmetries, 3D structure completion is also performed. Our approach significantly reduces drift through ”structural” loop closures and improves the accuracy of reconstructions in urban scenes.

In this paper, we present a framework for 6D absolute scale motion and structure estimation of a multi-camera system in challenging indoor environments. It operates in real-time and employs information from two cameras with non-overlapping fields of view. Monocular Visual Odometry supplying up-to-scale 6D motion information is carried out in each of the cameras, and the metric scale is recovered via a linear solution by imposing the known static transformation between both sensors. The redundancy in the motion estimates is finally exploited by a statistical fusion to an optimal 6D metric result. The proposed technique is robust to outliers and able to continuously deliver a reasonable measurement of the scale factor. The quality of the framework is demonstrated by a concise evaluation on indoor datasets, including a comparison to accurate ground truth data provided by an external motion tracking system.

Existing approaches to indoor scene understanding formulate the problem as a structured prediction task focusing on estimating the 3D bounding box which best describes the scene layout. Unfortunately, these approaches utilize high order potentials which are computationally intractable and rely on ad-hoc approximations for both learning and inference. In this paper we show that the potentials commonly used in the literature can be decomposed into pair-wise potentials by extending the concept of integral images to geometry. As a consequence no heuristic reduction of the search space is required. In practice, this results in large improvements in performance over the state-of-the-art, while being orders of magnitude faster.

In this work we present a unified view on Markov random fields and recently proposed continuous tight convex relaxations for multi-label assignment in the image plane. These relaxations are far less biased towards the grid geometry than Markov random fields. It turns out that the continuous methods are non-linear extensions of the local polytope MRF relaxation. In view of this result a better understanding of these tight convex relaxations in the discrete setting is obtained. Further, a wider range of optimization methods is now applicable to find a minimizer of the tight formulation. We propose two methods to improve the efficiency of minimization. One uses a weaker, but more efficient continuously inspired approach as initialization and gradually refines the energy where it is necessary. The other one reformulates the dual energy enabling smooth approximations to be used for efficient optimization. We demonstrate the utility of our proposed minimization schemes in numerical experiments.

The SFly project is an EU-funded project, with the goal to create a swarm of autonomous vision controlled micro aerial vehicles. The mission in mind is that a swarm of MAV's autonomously maps out an unknown environment, computes optimal surveillance positions and places the MAV's there and then locates radio beacons in this environment. The scope of the work includes contributions on multiple different levels ranging from theoretical foundations to hardware design and embedded programming. One of the contributions is the development of a new MAV, a hexacopter, equipped with enough processing power for onboard computer vision. A major contribution is the development of monocular visual SLAM that runs in real-time onboard of the MAV. The visual SLAM results are fused with IMU measurements and are used to stabilize and control the MAV. This enables autonomous flight of the MAV, without the need of a data link to a ground station. Within this scope novel analytical solutions for fusing IMU and vision measurements have been derived. In addition to the realtime local SLAM, an offline dense mapping process has been developed. For this the MAV's are equipped with a payload of a stereo camera system. The dense environment map is used to compute optimal surveillance positions for a swarm of MAV's. For this an optimiziation technique based on cognitive adaptive optimization has been developed. Finally, the MAV's have been equipped with radio transceivers and a method has been developed to locate radio beacons in the observed environment.

A wide range of robotic systems needs to estimate their rotation for diverse tasks like automatic control and stabilization, among many others. In regards of the limitations of traditional navigation equipments (like GPS and inertial sensors), this paper follows a vision approach based on the observation of vanishing points (VPs). Urban environments (outdoor as well as indoor) generally contain orthogonal VPs which constitutes an important constraint to fulfill in order to correctly acquire the structure of the scenes. In contrast to existing VP-based techniques, our method inherently enforces the orthogonality of the VPs by directly incorporating the orthogonality constraint into the model estimation step of the RANSAC procedure, which allows real-time applications. The model is estimated from only 3 lines, which corresponds to the theoretical minimal sampling for rotation estimation and constitutes our 3-line RANSAC. We also propose a 1-line RANSAC when the horizon plane is known. Our algorithm has been validated successfully on challenging real datasets.

In this paper, we describe our autonomous vision-based quadrotor MAV system which maps and explores unknown environments. All algorithms necessary for autonomous mapping and exploration run on-board the MAV. Using a front-looking stereo camera as the main exteroceptive sensor, our quadrotor achieves these capabilities with both the Vector Field Histogram+ (VFH+) algorithm for local navigation, and the frontier-based exploration algorithm. In addition, we implement the Bug algorithm for autonomous wall-following which could optionally be selected as the substitute exploration algorithm in sparse environments where the frontier-based exploration under-performs. We incrementally build a 3D global occupancy map on-board the MAV. The map is used by the VFH+ and frontier-based exploration in dense environments, and the Bug algorithm for wall-following in sparse environments. During the exploration phase, images from the front-looking camera are transmitted over Wi-Fi to the ground station. These images are input to a large-scale visual SLAM process running off-board on the ground station. SLAM is carried out with pose-graph optimization and loop closure detection using a vocabulary tree. We improve the robustness of the pose estimation by fusing optical flow and visual odometry. Optical flow data is provided by a customized downward-looking camera integrated with a microcontroller while visual odometry measurements are derived from the front-looking stereo camera. We verify our approaches with experimental results.

A high update rate of metric velocity values is crucial for a robust operation of navigation control loops of mobile robots such as micro aerial vehicles (MAVs). An efficient way for obtaining metric velocity of robots without external reference are image-based optical flow measurements, scaled with the distance between camera and the observed scene. However, since optical flow and stereo vision are computationally intensive tasks, metric optical flow calculations on embedded systems are typically only possible at limited frame rate. In this work, we therefore present an FPGA-based platform with the capability of calculating real-time metric optical flow at 127 frames per second and 376×240 resolution. Radial undistortion, image rectification, disparity estimation and optical flow calculation tasks are performed on a single FPGA without the need for external memory. The platform is perfectly suited for mobile robots or MAVs due to its low weight and low power consumption.

While finding the exact solution for the MAP inference problem is intractable for many real-world tasks, MAP LP relaxations have been shown to be very effective in practice. However, the most efficient methods that perform block coordinate descent can get stuck in sub-optimal points as they are not globally convergent. In this work we propose to augment these algorithms with an $\epsilon$-descent approach and present a method to efficiently optimize for a descent direction in the subdifferential using a margin-based extension of the Fenchel-Young duality theorem. Furthermore, the presented approach provides a methodology to construct a primal optimal solution from its dual optimal counterpart. We demonstrate the efficiency of the presented approach on spin glass models and protein interactions problems and show that our approach outperforms state-of-the-art solvers.

In sports broadcasts, networks consisting of pan-tilt-zoom (PTZ) cameras usually exhibit very wide baselines, making standard matching techniques for camera calibration very hard to apply. If, additionally, there is a lack of texture, finding corresponding image regions becomes almost impossible. However, such networks are often set up to observe dynamic scenes on a ground plane. Corresponding image trajectories produced by moving objects need to fulfill specific geometric constraints, which can be leveraged for camera calibration. We present a method which combines image trajectory matching with the self-calibration of rotating and zooming cameras, effectively reducing the remaining degrees of freedom in the matching stage to a 2D similarity transformation. Additionally, lines on the ground plane are used to improve the calibration. In the end, all extrinsic and intrinsic camera parameters are refined in a final bundle adjustment. The proposed algorithm was evaluated both qualitatively and quantitatively on four different soccer sequences.

With recent advances in mobile computing, the demand for visual localization or landmark identification on mobile devices is gaining interest. We advance the state of the art in this area by fusing two popular representations of street-level image data - facade-aligned and viewpoint-aligned - and show that they contain complementary information that can be exploited to significantly improve the recall rates on the city scale. We also improve feature detection in low contrast parts of the street-level data, and discuss how to incorporate priors on a user's position (e.g. given by noisy GPS readings or network cells), which previous approaches often ignore. Finally, and maybe most importantly, we present our results according to a carefully designed, repeatable evaluation scheme and make publicly available a set of 1.7 million images with ground truth labels, geotags, and calibration data, as well as a difficult set of cell phone query images. We provide these resources as a benchmark to facilitate further research in the area.

The paper proposes a vision based online mapping of large-scale environments. Our novel approach uses a hybrid representation of a fully metric Euclidean environment map and a topological map. This novel hybrid representation facilitates our scalable online hierarchical bundle adjustment approach. The proposed method achieves scalability by solving the local registration through embedding neighboring keyframes and landmarks into a Euclidean space. The global adjustment is performed on a segmentation of the keyframes and posed as the iterative optimization of the arrangement of keyframes in each segment and the arrangement of rigidly moving segments. The iterative global adjustment is performed concurrently with the local registration of the keyframes in a local map. Thus the map is always locally metric around the current location, and likely to be globally consistent. Loop closures are handled very efficiently benefiting from the topological nature of the map and overcoming the loss of the metric map properties as previous approaches. The effectiveness of the proposed method is demonstrated in real-time on various challenging video sequences.

In this paper we propose a distributed message-passing algorithm for inference in large scale graphical models. Our method can handle large problems efficiently by distributing and parallelizing the computation and memory requirements. The convergence and optimality guarantees of recently developed message-passing algorithms are preserved by introducing new types of consistency messages, sent between the distributed computers. We demonstrate the effectiveness of our approach in the task of stereo reconstruction from high-resolution imagery, and show that inference is possible with more than 200 labels in images larger than 10 MPixels.

How many people should you ask if you are not sure about your way? We provide an answer to this question for Random Forest classification. The presented method is based on the statistical formulation of confidence intervals and conjugate priors for binomial as well as multinomial distributions. We derive appealing decision rules to speed up the classification process by leveraging the fact that many samples can be clearly mapped to classes. Results on test data are provided, and we highlight the applicability of our method to a wide range of problems. The approach introduces only one non-heuristic parameter, that allows to trade-off accuracy and speed without any re-training of the classifier. The proposed method automatically adapts to the difficulty of the test data and makes classification significantly faster without deteriorating the accuracy.

We describe a generative model of the relationship between two images. The model is defined as a factored three-way Boltzmann machine, in which hidden variables collaborate to define the joint correlation matrix for image pairs. Modeling the joint distribution over pairs makes it possible to efficiently match images that are the same according to a learned measure of similarity. We apply the model to several face matching tasks, and show that it learns to represent the input images using task-specific basis functions. Matching performance is superior to previous similar generative models, including recent conditional models of transformations. We also show that the model can be used as a plug-in matching score to perform invariant classification.

This paper presents a novel approach for dense reconstruction from a single-view of a repetitive scene structure. Given an image and its detected repetition regions, we model the shape recovery as the dense pixel correspondences within a single image. The correspondences are represented by an interval map that tells the distance of each pixel to its matched pixels within the single image. In order to obtain dense repetitive structures, we develop a new repetition constraint that penalizes the inconsistency between the repetition intervals of the dynamically corresponding pixel pairs. We deploy a graph-cut to balance between the high-level constraint of geometric repetition and the low-level constraints of photometric consistency and spatial smoothness. We demonstrate the accurate reconstruction of dense 3D repetitive structures through a variety of experiments, which prove the robustness of our approach to outliers such as structure variations, illumination changes, and occlusions.

In geometric computer vision, the structure from motion (SfM) problem can be formulated as a optimization problem with a rank constraint. It is well known that the trace norm of a matrix can act as a convex proxy for a low rank constraint. Hence, in recent work [7], the trace-norm relaxation has been applied to the SfM problem. However, SfM problems often exhibit a certain structure, for example a smooth camera path. Unfortunately, the trace norm relaxation can not make use of this additional structure. This observation motivates the main contribution of this paper. We present the so-called generalized trace norm which allows to encode prior knowledge about a specific problem into a convex regularization term which enforces a low rank solution while at the same time taking the problem structure into account. While deriving the generalized trace norm and stating its different formulations, we draw interesting connections to other fields, most importantly to the field of compressive sensing. Even though the generalized trace norm is a very general concept with a wide area of potential applications we are ultimately interested in applying it to SfM problems. Therefore, we also present an efficient algorithm to optimize the resulting generalized trace norm regularized optimization problems. Results show that the generalized trace norm indeed achieves its goals in providing a problem-dependent regularization.

In this paper, we propose an efficient technique to detect changes in the geometry of an urban environment using some images observing its current state. The proposed method can be used to significantly optimize the process of updating the 3D model of a city changing over time, by restricting this process to only those areas where changes are detected. With this application in mind, we designed our algorithm to specifically detect only structural changes in the environment, ignoring any changes in its appearance, and ignoring also all the changes which are not relevant for update purposes, such as cars, people etc. As a by-product, the algorithm also provides a coarse geometry of the detected changes. The performance of the proposed method was tested on four different kinds of urban environments and compared with two alternative techniques.

This paper presents a novel system to estimate body pose configuration from a single depth map. It combines both pose detection and pose refinement. The input depth map is matched with a set of pre-captured motion exemplars to generate a body configuration estimation, as well as semantic labeling of the input point cloud. The initial estimation is then refined by directly fitting the body configuration with the observation (e.g., the input depth). In addition to the new system architecture, our other contributions include modifying a point cloud smoothing technique to deal with very noisy input depth maps, a point cloud alignment and pose search algorithm that is view-independent and efficient. Experiments on a public dataset show that our approach achieves significantly higher accuracy than previous state-of-art methods.

We present a novel stereo-based obstacle avoidance system on a vision-guided micro air vehicle (MAV) that is capable of fully autonomous maneuvers in unknown and dynamic environments. All algorithms run exclusively on the vehicle's on-board computer, and at high frequencies that allow the MAV to react quickly to obstacles appearing in its flight trajectory. Our MAV platform is a quadrotor aircraft equipped with an inertial measurement unit and two stereo rigs. An obstacle mapping algorithm processes stereo images, producing a 3D map representation of the environment; at the same time, a dynamic anytime path planner plans a collision-free path to a goal point.

Bundle adjustment (BA) which produces highly accurate results for visual Simultaneous Localization and Mapping (SLAM) could not be used for Micro-Aerial Vehicles (MAVs) with limited processing power because of its O(N3) complexity. We observed that a consistent ground plane often exists for MAVs flying in both the indoor and outdoor urban environments. Therefore, in this paper, we propose a visual SLAM algorithm that make use of the plane constraint to reduce the complexity of BA. The reduction of complexity is achieved by refining only the current camera pose and most recent map points with BA that minimizes the reprojection errors and perpendicular distances between the most recent map points and the best fit plane with all the pre-existing map points. As a result, our algorithm is approximately constant time since the number of current camera pose and most recent map points remain approximately constant. In addition, the minimization of the perpendicular distances between the plane and map points would enforce consistency between the reconstructed map points and the actual ground plane.

We provide a novel hardware and software system for micro air vehicles (MAV) that allows high-speed, low-latency onboard image processing. It uses up to four cameras in parallel on a miniature rotary wing platform. The MAV navigates based on onboard processed computer vision in GPS-denied in- and outdoor environments. It can process in parallel images and inertial measurement information from multiple cameras for multiple purposes (localization, pattern recognition, obstacle avoidance) by distributing the images on a central, low-latency image hub. Furthermore the system can utilize low-bandwith radio links for communication and is designed and optimized to scale to swarm use. Experimental results show successful flight with a range of onboard computer vision algorithms, including localization, obstacle avoidance and pattern recognition.

We present a method to reconstruct indoor environments from stereo image pairs, suitable for the navigation of robots. To enable a robot to navigate solely using visual cues it receives from a stereo camera, the depth information needs to be extracted from the image pairs and combined into a common representation. The initially determined raw depthmaps are fused into a two level heightmap representation which contains a floor and a ceiling height level. To reduce the noise in the height maps we employ a total variation regularized energy functional.With this 2.5D representation of the scene the computational complexity of the energy optimization is reduced by one dimension in contrast to other fusion techniques that work on the full 3D space such as volumetric fusion. While we show only results for indoor environments the approach can be extended to generate heightmaps for outdoor environments.

In this paper, we proposed a method to recognize complex human daily activities including body activities and hand gestures simultaneously in an indoor environment. Three wearable motion sensors are attached to the right thigh, the waist, and the right hand of a person, while an optical motion capture system is used to obtain his/her location information. A three-level dynamic Bayesian network is implemented to model the intra-temporal and inter-temporal constraints among the location, body activity and hand gesture. The body activity and hand gesture are estimated using a Bayesian filter and the short-time Viterbi algorithm, which reduces the storage memory and the computational complexity. We conducted experiments in a mock apartment environment and the obtained results showed the effectiveness and accuracy of our algorithms.

In this paper, we present our RS-SLAM algorithm for monocular camera where the proposal distribution is derived from the 5-point RANSAC algorithm and image feature measurement uncertainties instead of using the easily violated constant velocity model. We propose to do another RANSAC sampling within all the inliers that have the best RANSAC score to check for inlier misclassifications in the original correspondences and use all the hypotheses generated from these consensus sets in the proposal distribution. This is to mitigate data association errors (inlier misclassifications) caused by the observation that the consensus set from RANSAC that yields the highest score might not, in practice, contain all the true inliers due to noise on the feature measurements. Hypotheses which are less probable will eventually be eliminated in the particle filter resampling process. We also show in this paper that our monocular approach can be easily extended for stereo camera. Experimental results validate the potential of our approach.

Real-world camera networks are often characterized by very wide baselines covering a wide range of viewpoints. We describe a method not only calibrating each camera sequence added to the system automatically, but also taking advantage of multi-view correspondences to make the entire calibration framework more robust. Novel camera sequences can be seamlessly integrated into the system at any time, adding to the robustness of future computations. One of the challenges consists in establishing correspondences between cameras. Initializing a bag of features from a calibrated frame, correspondences between cameras are established in a two-step procedure. First, affine invariant features of camera sequences are warped into a common coordinate frame and a coarse matching is obtained between the collected features and the incrementally built and updated bag of features. This allows us to warp images to a common view. Second, scale invariant features are extracted from the warped images. This leads to both more numerous and more accurate correspondences. Finally, the parameters are optimized in a bundle adjustment. Adding the feature descriptors and the optimized 3D positions to the bag of features, we obtain a feature-based scene abstraction, allowing for the calibration of novel sequences and the correction of drift in single-view calibration tracking. We demonstrate that our approach can deal with wide baselines. Novel sequences can seamlessly be integrated in the calibration framework.

. Kaiming He, Jian Sun, and Xiaoou Tang, “Single Image Haze Removal Using Dark Channel Prior,” winner of the Best Paper Award (sponsored by Microsoft). . Anat Levin, Yair Weiss, Fredo Durand, and Bill Freeman, “Understanding and evaluating blind deconvolution algorithms,” winner of the Best Paper-Honorable Mention Award (sponsored by Honeywell). . Ce Liu, Jenny Yuen, and Antonio Torralba, “Nonparametric Scene Parsing: Label Transfer via Dense Scene Alignment,” winner of the Best Student Paper Award (sponsored by MERL). . Olivier Duchenne, Francis Bach, In So Kweon, and Jean Ponce, “A Tensor-Based Algorithm for HighOrder Graph Matching,” winner of the Best Student Paper-Honorable Mention Award (sponsored by Hewlett-Packard). CVPR 2009 received 1,464 complete submissions by the 20 November 2008 deadline. We worked with 46 Area Chairs, who are well-respected members of the computer vision community, and 749 reviewers to select 61 papers as Orals and 322 papers as Posters. Twenty of these accepted papers were recommended to the Awards Committee for consideration. The Awards Committee consisted of five senior members of the vision community, three of whom were Area Chairs. The committee members had no conflicts with the candidate papers. The four award papers were selected after three phases of reviewing. These award papers were presented in the only singletrack session of the main conference. The authors of the award-winning papers and honorable mentions were invited to submit an extended version of their paper as a journal submission to TPAMI. The papers were reviewed by expert reviewers in the field, following the usual TPAMI procedure. All of the papers were accepted, in most cases after relatively minor revisions. CVPR 2009 also presented the Longuet-Higgins Prize for fundamental contributions in computer vision that have withstood the test of time. The awards committee that selected the CVPR 2009 awards was also asked to assist in the selection of this award. The prize (sponsored by IBM) was awarded to the following two papers that were published in CVPR 1999:

Given a set of algorithms, which one(s) should you apply to, i) compute optical flow, or ii) perform feature matching? Would looking at the sequence in question help you decide? It is unclear if even a person with intimate knowledge of all the different algorithms and access to the sequence itself could predict which one to apply. Our hypothesis is that the most suitable algorithm can be chosen for each video automatically, through supervised training of a classifier. The classifier treats the different algorithms as black-box alternative “classes,” and predicts when each is best because of their respective performances on training examples where ground truth flow was available. Our experiments show that a simple Random Forest classifier is predictive of algorithm-suitability. The automatic feature selection makes use of both our spatial and temporal video features. We find that algorithm-suitability can be determined per-pixel, capitalizing on the heterogeneity of appearance and motion within a video. We demonstrate our learned region segmentation approach quantitatively using four available flow algorithms, on both known and novel image sequences with ground truth flow. We achieve performance that often even surpasses that of the one best algorithm at our disposal.

Piecewise planar models for stereo have recently become popular for modeling indoor and urban outdoor scenes. The strong planarity assumption overcomes the challenges presented by poorly textured surfaces, and results in low complexity 3D models for rendering, storage, and transmission. However, such a model performs poorly in the presence of non-planar objects, for example, bushes, trees, and other clutter present in many scenes. We present a stereo method capable of handling more general scenes containing both planar and non-planar regions. Our proposed technique segments an image into piecewise planar regions as well as regions labeled as non-planar. The non-planar regions are modeled by the results of a standard multi-view stereo algorithm. The segmentation is driven by multi-view photoconsistency as well as the result of a color-and texture-based classifier, learned from hand-labeled planar and non-planar image regions. Additionally our method links and fuses plane hypotheses across multiple overlapping views, ensuring a consistent 3D reconstruction over an arbitrary number of images. Using our system, we have reconstructed thousands of frames of street-level video. Results show our method successfully recovers piecewise planar surfaces alongside general 3D surfaces in challenging scenes containing large buildings as well as residential houses.