Event cameras in motion tend to detect object boundaries or texture edges, which produce lines of brightness changes, especially in man-made environments. While lines can constitute a robust intermediate representation that is consistently observed, the sparse nature of lines may lead to drastic deterioration with minor estimation errors. Only a few previous works, often accompanied by additional sensors, utilize lines to compensate for the severe domain discrepancies of event sensors along with unpredictable noise characteristics. We propose a method that can stably extract tracks of varying appearances of lines using a clever algorithmic process that observes multiple representations from various time slices of events, compensating for potential adversaries within the event data. We then propose geometric cost functions that can refine the 3D line maps and camera poses, eliminating projective distortions and depth ambiguities. The 3D line maps are highly compact and can be equipped with our proposed cost function, which can be adapted for any observations that can detect and extract line structures or projections of them, including 3D point cloud maps or image observations. We demonstrate that our formulation is powerful enough to exhibit a significant performance boost in event-based mapping and pose refinement across diverse datasets, and can be flexibly applied to multimodal scenarios. Our results confirm that the proposed line-based formulation is a robust and effective approach for the practical deployment of event-based perceptual modules. Project page

We propose scene-adaptive strategies to efficiently allocate representation capacity for generating immersive experiences of indoor environments from incomplete observations. Indoor scenes with multiple rooms often exhibit irregular layouts with varying complexity, containing clutter, occlusion, and flat walls. We maximize the utilization of limited resources with guidance from geometric priors, which are often readily available after pre-processing stages. We record observation statistics on the estimated geometric scaffold and guide the optimal placement of bases, which greatly improves upon the uniform basis arrangements adopted by previous scalable scene representations. We also suggest scene-adaptive virtual viewpoints to compensate for geometric deficiencies inherent in view configurations in the input trajectory and impose the necessary regularization. We present a comprehensive analysis and discussion regarding rendering quality and memory requirements in several large-scale indoor scenes, demonstrating significant enhancements compared to baselines that employ regular placements.

Understanding 3D motion from videos presents inherent challenges due to the diverse types of movement, ranging from rigid and deformable objects to articulated structures. To overcome this, we propose Liv3Stroke, a novel approach for abstracting objects in motion with deformable 3D strokes. The detailed movements of an object may be represented by unstructured motion vectors or a set of motion primitives using a pre-defined articulation from a template model. Just as a free-hand sketch can intuitively visualize scenes or intentions with a sparse set of lines, we utilize a set of parametric 3D curves to capture a set of spatially smooth motion elements for general objects with unknown structures. We first extract noisy, 3D point cloud motion guidance from video frames using semantic features, and our approach deforms a set of curves to abstract essential motion features as a set of explicit 3D representations. Such abstraction enables an understanding of prominent components of motions while maintaining robustness to environmental factors. Our approach allows direct analysis of 3D object movements from video, tackling the uncertainty that typically occurs when translating real-world motion into recorded footage. The project page is accessible via: https://jaeah.me/liv3stroke_web.

Recent advances in motion diffusion models have led to remarkable progress in diverse motion generation tasks, including text-to-motion synthesis. However, existing approaches represent motions as dense frame sequences, requiring the model to process redundant or less informative frames. The processing of dense animation frames imposes significant training complexity, especially when learning intricate distributions of large motion datasets even with modern neural architectures. This severely limits the performance of generative motion models for downstream tasks. Inspired by professional animators who mainly focus on sparse keyframes, we propose a novel diffusion framework explicitly designed around sparse and geometrically meaningful keyframes. Our method reduces computation by masking non-keyframes and efficiently interpolating missing frames. We dynamically refine the keyframe mask during inference to prioritize informative frames in later diffusion steps. Extensive experiments show that our approach consistently outperforms state-of-the-art methods in text alignment and motion realism, while also effectively maintaining high performance at significantly fewer diffusion steps. We further validate the robustness of our framework by using it as a generative prior and adapting it to different downstream tasks.

Recent works on dynamic 3D neural field reconstruction assume the input from synchronized multi-view videos whose poses are known. The input constraints are often not satisfied in real-world setups, making the approach impractical. We show that unsynchronized videos from unknown poses can generate dynamic neural fields as long as the videos capture human motion. Humans are one of the most common dynamic subjects captured in videos, and their shapes and poses can be estimated using state-of-the-art libraries. While noisy, the estimated human shape and pose parameters provide a decent initialization point to start the highly non-convex and under-constrained problem of training a consistent dynamic neural representation. Given the shape and pose parameters of humans in individual frames, we formulate methods to calculate the time offsets between videos, followed by camera pose estimations that analyze the 3D joint positions. Then, we train the dynamic neural fields employing multiresolution grids while we concurrently refine both time offsets and camera poses. The setup still involves optimizing many parameters; therefore, we introduce a robust progressive learning strategy to stabilize the process. Experiments show that our approach achieves accurate spatio-temporal calibration and high-quality scene reconstruction in challenging conditions.

Creating expressive character animations is labor-intensive, requiring intricate manual adjustment of animators across space and time. Previous works on controllable motion generation often rely on a predefined set of dense spatio-temporal specifications (e.g., dense pelvis trajectories with exact per-frame timing), limiting practicality for animators. To process high-level intent and intuitive control in diverse scenarios, we propose a practical controllable motions synthesis framework that respects sparse and flexible keyjoint signals. Our approach employs a decomposed diffusion-based motion synthesis framework that first synthesizes keyjoint movements from sparse input control signals and then synthesizes full-body motion based on the completed keyjoint trajectories. The low-dimensional keyjoint movements can easily adapt to various control signal types, such as end-effector position for diverse goal-driven motion synthesis, or incorporate functional constraints on a subset of keyjoints. Additionally, we introduce a time-agnostic control formulation, eliminating the need for frame-specific timing annotations and enhancing control flexibility. Then, the shared second stage can synthesize a natural whole-body motion that precisely satisfies the task requirement from dense keyjoint movements. We demonstrate the effectiveness of sparse and flexible keyjoint control through comprehensive experiments on diverse datasets and scenarios.

Modeling human-scene interactions (HSI) is essential for understanding and simulating everyday human behaviors. Recent approaches utilizing generative modeling have made progress in this domain; however, they are limited in controllability and flexibility for real-world applications. To address these challenges, we propose reformulating the HSI modeling problem as Scene-aware Motion In-betweening - a more tractable and practical task. We introduce SceneMI, a framework that supports several practical applications, including keyframe-guided character animation in 3D scenes and enhancing the motion quality of imperfect HSI data. SceneMI employs dual scene descriptors to comprehensively encode global and local scene context. Furthermore, our framework leverages the inherent denoising nature of diffusion models to generalize on noisy keyframes. Experimental results demonstrate SceneMI's effectiveness in scene-aware keyframe in-betweening and generalization to the real-world GIMO dataset, where motions and scenes are acquired by noisy IMU sensors and smartphones. We further showcase SceneMI's applicability in HSI reconstruction from monocular videos.

In this work, we propose a framework that creates a lively virtual dynamic scene with contextual motions of multiple humans. Generating multi-human contextual motion requires holistic reasoning over dynamic relationships among human-human and human-scene interactions. We adapt the power of a large language model (LLM) to digest the contextual complexity within textual input and convert the task into tangible subproblems such that we can generate multi-agent behavior beyond the scale that was not considered before. Specifically, our event generator formulates the temporal progression of a dynamic scene into a sequence of small events. Each event calls for a well-defined motion involving relevant characters and objects. Next, we synthesize the motions of characters at positions sampled based on spatial guidance. We employ a high-level module to deliver scalable yet comprehensive context, translating events into relative descriptions that enable the retrieval of precise coordinates. As the first to address this problem at scale and with diversity, we offer a benchmark to assess diverse aspects of contextual reasoning. Benchmark results and user studies show that our framework effectively captures scene context with high scalability. The code and benchmark, along with result videos, are available at our project page: https://rms0329.github.io/Event-Driven-Storytelling/.

While conventional data are represented as discrete vectors, Implicit Neural Representations (INRs) utilize neural networks to represent data points as continuous functions. By incorporating a shared network that maps latent vectors to individual functions, one can model the distribution of functional data, which has proven effective in many applications, such as learning 3D shapes, surface reflectance, and operators. However, the infinite-dimensional nature of these representations makes them prone to overfitting, necessitating sufficient regularization. Naïve regularization methods -- those commonly used with discrete vector representations -- may enforce smoothness to increase robustness but result in a loss of data fidelity due to improper handling of function coordinates. To overcome these challenges, we start by interpreting the mapping from latent variables to INRs as a parametrization of a Riemannian manifold. We then recognize that preserving geometric quantities -- such as distances and angles -- between the latent space and the data manifold is crucial. As a result, we obtain a manifold with minimal intrinsic curvature, leading to robust representations while maintaining high-quality data fitting. Our experiments on various data modalities demonstrate that our method effectively discovers a well-structured latent space, leading to robust data representations even for challenging datasets, such as those that are small or noisy.

Abstract We present a versatile latent representation that enables physically simulated character to efficiently utilize motion priors. To build a powerful motion embedding that is shared across multiple tasks, the physics controller should employ rich latent space that is easily explored and capable of generating high‐quality motion. We propose integrating continuous and discrete latent representations to build a versatile motion prior that can be adapted to a wide range of challenging control tasks. Specifically, we build a discrete latent model to capture distinctive posterior distribution without collapse, and simultaneously augment the sampled vector with the continuous residuals to generate high‐quality, smooth motion without jittering. We further incorporate Residual Vector Quantization, which not only maximizes the capacity of the discrete motion prior, but also efficiently abstracts the action space during the task learning phase. We demonstrate that our agent can produce diverse yet smooth motions simply by traversing the learned motion prior through unconditional motion generation. Furthermore, our model robustly satisfies sparse goal conditions with highly expressive natural motions, including head‐mounted device tracking and motion in‐betweening at irregular intervals, which could not be achieved with existing latent representations.

We consider the problem of client-server localization, where edge device users communicate visual data with the service provider for locating oneself against a pre-built 3D map. This localization paradigm is a crucial component for location-based services in AR/VR or mobile applications, as it is not trivial to store large-scale 3D maps and process fast localization on resource-limited edge devices. Nevertheless, conventional client-server localization systems possess numerous challenges in computational efficiency, robustness, and privacy-preservation during data transmission. Our work aims to jointly solve these challenges with a localization pipeline based on event cameras. By using event cameras, our system consumes low energy and maintains small memory bandwidth. Then during localization, we propose applying event-to-image conversion and leverage mature image-based localization, which achieves robustness even in low-light or fast-moving scenes. To further enhance privacy protection, we introduce privacy protection techniques at two levels. Network level protection aims to hide the entire user's view in private scenes using a novel split inference approach, while sensor level protection aims to hide sensitive user details such as faces with light-weight filtering. Both methods involve small client-side computation and localization performance loss, while significantly mitigating the feeling of insecurity as revealed in our user study. We thus project our method to serve as a building block for practical location-based services using event cameras.

We introduce a lightweight and accurate localization method that only utilizes the geometry of 2D-3D lines. Given a pre-captured 3D map, our approach localizes a panorama image, taking advantage of the holistic 360° view. The system mitigates potential privacy breaches or domain discrepancies by avoiding trained or hand-crafted visual descriptors. However, as lines alone can be ambiguous, we express distinctive yet compact spatial contexts from relationships between lines, namely the dominant directions of parallel lines and the intersection between non-parallel lines. The resulting representations are efficient in processing time and memory compared to conventional visual descriptor-based methods. Given the groups of dominant line directions and their intersections, we accelerate the search process to test thousands of pose candidates in less than a millisecond without sacrificing accuracy. We empirically show that the proposed 2D-3D matching can localize panoramas for challenging scenes with similar structures, dramatic domain shifts or illumination changes. Our fully geometric approach does not involve extensive parameter tuning or neural network training, making it a practical algorithm that can be readily deployed in the real world. Project page including the code is available through this link: https://82magnolia.github.io/fgpl/.

We aim to generate fine-grained 3D geometry from large-scale sparse LiDAR scans, abundantly captured by autonomous vehicles (AV). Contrary to prior work on AV scene completion, we aim to extrapolate fine geometry from unlabeled and beyond spatial limits of LiDAR scans, taking a step towards generating realistic, high-resolution simulation-ready 3D street environments. We propose hierarchical Generative Cellular Automata (hGCA), a spatially scalable conditional 3D generative model, which grows geometry recursively with local kernels following [46, 47], in a coarse-to-fine manner, equipped with a light-weight planner to induce global consistency. Experiments on synthetic scenes show that hGCA generates plausible scene geometry with higher fidelity and completeness compared to state-of-the-art baselines. Our model generalizes strongly from sim-to-real, qualitatively outperforming baselines on the Waymo-open dataset. We also show anecdotal evidence of the ability to create novel objects from real-world geometric cues even when trained on limited synthetic content. More results and details can be found on our project page.

While free-hand sketching has long served as an efficient representation to convey characteristics of an object, they are often subjective, deviating significantly from realistic representations. Moreover, sketches are not consistent for arbitrary viewpoints, making it hard to catch 3D shapes. We propose 3Dooole, generating descriptive and view-consistent sketch images given multi-view images of the target object. Our method is based on the idea that a set of 3D strokes can efficiently represent 3D structural information and render view-consistent 2D sketches. We express 2D sketches as a union of view-independent and view-dependent components. 3D cubic Bézier curves indicate view-independent 3D feature lines, while contours of superquadrics express a smooth outline of the volume of varying viewpoints. Our pipeline directly optimizes the parameters of 3D stroke primitives to minimize perceptual losses in a fully differentiable manner. The resulting sparse set of 3D strokes can be rendered as abstract sketches containing essential 3D characteristic shapes of various objects. We demonstrate that 3Doodle can faithfully express concepts of the original images compared with recent sketch generation approaches. 1

We present EgoNeRF, a practical solution to reconstruct large-scale real-world environments for VR assets. Given a few seconds of casually captured 360 video, EgoNeRF can efficiently build neural radiance fields. Motivated by the recent acceleration of NeRF using feature grids, we adopt spherical coordinate instead of conventional Cartesian coordinate. Cartesian feature grid is inefficient to represent large-scale unbounded scenes because it has a spatially uniform resolution, regardless of distance from viewers. The spherical parameterization better aligns with the rays of egocentric images, and yet enables factorization for performance enhancement. However, the naive spherical grid suffers from singularities at two poles, and also cannot represent unbounded scenes. To avoid singularities near poles, we combine two balanced grids, which results in a quasi-uniform angular grid. We also partition the radial grid exponentially and place an environment map at infinity to represent unbounded scenes. Furthermore, with our resampling technique for grid-based methods, we can increase the number of valid samples to train NeRF volume. We extensively evaluate our method in our newly introduced synthetic and real-world egocentric 360 video datasets, and it consistently achieves state-of-the-art performance.

We propose Text2Scene, a method to automatically create realistic textures for virtual scenes composed of multiple objects. Guided by a reference image and text descriptions, our pipeline adds detailed texture on labeled 3D geometries in the room such that the generated colors respect the hierarchical structure or semantic parts that are often composed of similar materials. Instead of applying flat stylization on the entire scene at a single step, we obtain weak semantic cues from geometric segmentation, which are further clarified by assigning initial colors to segmented parts. Then we add texture details for individual objects such that their projections on image space exhibit feature embedding aligned with the embedding of the input. The decomposition makes the entire pipeline tractable to a moderate amount of computation resources and memory. As our framework utilizes the existing resources of image and text embedding, it does not require dedicated datasets with high-quality textures designed by skillful artists. To the best of our knowledge, it is the first practical and scalable approach that can create detailed and realistic textures of the desired style that maintain structural context for scenes with multiple objects.

The exact 3D dynamics of the human body provides crucial evidence to analyze the consequences of the physical interaction between the body and the environment, which can eventually assist everyday activities in a wide range of applications. However, optimizing for 3D configurations from image observation requires a significant amount of computation, whereas real-world 3D measurements often suffer from noisy observation or complex occlusion. We resolve the challenge by learning a latent distribution representing strong temporal priors. We use a conditional variational autoencoder (CVAE) architecture with a transformer to train the motion priors with a large-scale motion dataset. Then our feature follower effectively aligns the feature spaces of noisy, partial observation with the necessary input for pre-trained motion priors, and quickly recovers a complete mesh sequence of motion. We demonstrate that the transformer-based autoencoder can collect necessary spatio-temporal correlations robust to various adversaries, such as missing temporal frames, or noisy observation under severe occlusion. Our framework is general and can be applied to recover the full 3D dynamics of other subjects with parametric representations.

We introduce LDL, a fast and robust algorithm that localizes a panorama to a 3D map using line segments. LDL focuses on the sparse structural information of lines in the scene, which is robust to illumination changes and can potentially enable efficient computation. While previous line-based localization approaches tend to sacrifice accuracy or computation time, our method effectively observes the holistic distribution of lines within panoramic images and 3D maps. Specifically, LDL matches the distribution of lines with 2D and 3D line distance functions, which are further decomposed along principal directions of lines to increase the expressiveness. The distance functions provide coarse pose estimates by comparing the distributional information, where the poses are further optimized using conventional local feature matching. As our pipeline solely leverages line geometry and local features, it does not require costly additional training of line-specific features or correspondence matching. Nevertheless, our method demonstrates robust performance on challenging scenarios including object layout changes, illumination shifts, and large-scale scenes, while exhibiting fast pose search terminating within a matter of milliseconds. We thus expect our method to serve as a practical solution for line-based localization, and complement the well-established point-based paradigm. The code for LDL is available through the following link: https://github.com/82magnolia/panoramic-localization.

The absolute depth values of surrounding environments provide crucial cues for various assistive technologies, such as localization, navigation, and 3D structure estimation. We propose that accurate depth estimated from panoramic images can serve as a powerful and light-weight input for a wide range of downstream tasks requiring 3D information. While panoramic images can easily capture the surrounding context from commodity devices, the estimated depth shares the limitations of conventional image-based depth estimation; the performance deteriorates under large domain shifts and the absolute values are still ambiguous to infer from 2D observations. By taking advantage of the holistic view, we mitigate such effects in a self-supervised way and fine-tune the network with geometric consistency during the test phase. Specifically, we construct a 3D point cloud from the current depth prediction and project the point cloud at various viewpoints or apply stretches on the current input image to generate synthetic panoramas. Then we minimize the discrepancy of the 3D structure estimated from synthetic images without collecting additional data. We empirically evaluate our method in robot navigation and map-free localization where our method shows large performance enhancements. Our calibration method can therefore widen the applicability under various external conditions, serving as a key component for practical panorama-based machine vision systems.

We present Ev-NeRF, a Neural Radiance Field derived from event data. While event cameras can measure subtle brightness changes in high frame rates, the measurements in low lighting or extreme motion suffer from significant domain discrepancy with complex noise. As a result, the performance of event-based vision tasks does not transfer to challenging environments, where the event cameras are expected to thrive over normal cameras. We find that the multi-view consistency of NeRF provides a powerful self-supervision signal for eliminating spurious measurements and extracting the consistent underlying structure despite highly noisy input. Instead of posed images of the original NeRF, the input to Ev-NeRF is the event measurements accompanied by the movements of the sensors. Using the loss function that reflects the measurement model of the sensor, Ev-NeRF creates an integrated neural volume that summarizes the unstructured and sparse data points captured for about 2-4 seconds. The generated neural volume can also produce intensity images from novel views with reasonable depth estimates, which can serve as a high-quality input to various vision-based tasks. Our results show that Ev-NeRF achieves competitive performance for intensity image reconstruction under extreme noise and high-dynamic-range imaging.

Abstract We propose IBL‐NeRF, which decomposes the neural radiance fields (NeRF) of large‐scale indoor scenes into intrinsic components. Recent approaches further decompose the baked radiance of the implicit volume into intrinsic components such that one can partially approximate the rendering equation. However, they are limited to representing isolated objects with a shared environment lighting, and suffer from computational burden to aggregate rays with Monte Carlo integration. In contrast, our prefiltered radiance field extends the original NeRF formulation to capture the spatial variation of lighting within the scene volume, in addition to surface properties. Specifically, the scenes of diverse materials are decomposed into intrinsic components for rendering, namely, albedo, roughness, surface normal, irradiance, and prefiltered radiance. All of the components are inferred as neural images from MLP, which can model large‐scale general scenes. Especially the prefiltered radiance effectively models the volumetric light field, and captures spatial variation beyond a single environment light. The prefiltering aggregates rays in a set of predefined neighborhood sizes such that we can replace the costly Monte Carlo integration of global illumination with a simple query from a neural image. By adopting NeRF, our approach inherits superior visual quality and multi‐view consistency for synthesized images as well as the intrinsic components. We demonstrate the performance on scenes with complex object layouts and light configurations, which could not be processed in any of the previous works.

Abstract The advancements of the Neural Radiance Field (NeRF) and its variants have demonstrated remarkable capabilities in generating photo‐realistic novel views from a small set of input images. While recent works suggest various techniques and model architectures that enhance speed or reconstruction quality, little attention is paid to exploring the RGB color space of input images. In this paper, we propose a universal color transform module that can maximally harness the captured evidence for the neural networks at hand. The color transform module utilizes an encoder‐decoder framework that maps the RGB color space into a new latent space, enhancing the expressiveness of the input domain. We attach the encoder and the decoder at the input and output of a NeRF model of choice, respectively, and jointly optimize them to maintain the cycle consistency of the proposed transform, in addition to minimizing the reconstruction errors in the feature domain. Our comprehensive experiments demonstrate that the learned color space can significantly improve the quality of reconstructions compared to the conventional RGB representation. Its benefits are particularly pronounced in challenging scenarios characterized by low‐light environments and scenes with low‐textured regions. The proposed color transform pushes the boundaries of limitations in the input domain and offers a promising avenue for advancing the reconstruction capabilities of various neural representations. Source code is available at https://github.com/sangminkim-99/ColorTransformModule .

We present Neural Marionette, an unsupervised approach that discovers the skeletal structure from a dynamic sequence and learns to generate diverse motions that are consistent with the observed motion dynamics. Given a video stream of point cloud observation of an articulated body under arbitrary motion, our approach discovers the unknown low-dimensional skeletal relationship that can effectively represent the movement. Then the discovered structure is utilized to encode the motion priors of dynamic sequences in a latent structure, which can be decoded to the relative joint rotations to represent the full skeletal motion. Our approach works without any prior knowledge of the underlying motion or skeletal structure, and we demonstrate that the discovered structure is even comparable to the hand-labeled ground truth skeleton in representing a 4D sequence of motion. The skeletal structure embeds the general semantics of possible motion space that can generate motions for diverse scenarios. We verify that the learned motion prior is generalizable to the multi-modal sequence generation, interpolation of two poses, and motion retargeting to a different skeletal structure.

We introduce Ev-TTA, a simple, effective test-time adaptation algorithm for event-based object recognition. While event cameras are proposed to provide measurements of scenes with fast motions or drastic illumination changes, many existing event-based recognition algorithms suffer from performance deterioration under extreme conditions due to significant domain shifts. Ev-TTA mitigates the severe domain gaps by fine-tuning the pre-trained classifiers during the test phase using loss functions inspired by the spatio-temporal characteristics of events. Since the event data is a temporal stream of measurements, our loss function enforces similar predictions for adjacent events to quickly adapt to the changed environment online. Also, we utilize the spatial correlations between two polarities of events to handle noise under extreme illumination, where different polarities of events exhibit distinctive noise distributions. Ev-TTA demonstrates a large amount of performance gain on a wide range of event-based object recognition tasks without extensive additional training. Our formulation can be successfully applied regardless of input representations and further extended into regression tasks. We expect Ev-TTA to provide the key technique to deploy event-based vision algorithms in challenging real-world applications where significant domain shift is inevitable.

In this paper, we introduce a mask-based grasping method that discerns multiple objects within the scene regard-less of transparency or specularity and finds the optimal grasp position avoiding clutter. Conventional vision-based robotic grasping approaches often fail to extend to the scenes containing transparent objects due to their different visual appearance. To handle the different visual characteristics, we first segment both transparent and opaque objects into instance masks, which serve as the domain-agnostic intermediate representation of both object types, using a neural network. While there exists no labelled training dataset that strongly represents both object types, we overcome the limitation by augmenting transparent objects on an existing large-scale dataset. Then, given the object instance masks, our method selects the top K discrete masks and robustly estimates grasp poses avoiding clutter. Through experiments, we verify that the instance masks are light-weight yet provide sufficient information for vision-based grasping agnostic of various appearances. On an unseen real-world test environment with complex objects, our method substantially outperforms previous methods without fine-tuning.

We propose a light-weight, self-supervised adaptation for a visual navigation agent to generalize to unseen environment. Given an embodied agent trained in a noiseless environment, our objective is to transfer the agent to a noisy environment where actuation and odometry sensor noise is present. Our method encourages the agent to maximize the consistency between the global maps generated at different time steps in a round-trip trajectory. The proposed task is completely self-supervised, not requiring any supervision from ground-truth pose data or explicit noise model. In addition, optimization of the task objective is extremely light-weight, as training terminates within a few minutes on a commodity GPU. Our experiments show that the proposed task helps the agent to successfully transfer to new, noisy environments. The transferred agent exhibits improved localization and mapping accuracy, further leading to enhanced performance in downstream visual navigation tasks. Moreover, we demonstrate test-time adaptation with our self-supervised task to show its potential applicability in real-world deployment.

We present GATSBI, a generative model that can transform a sequence of raw observations into a structured latent representation that fully captures the spatiotemporal context of the agent’s actions. In vision-based decision making scenarios, an agent faces complex high-dimensional observations where multiple entities interact with each other. The agent requires a good scene representation of the visual observation that discerns essential components and consistently propagates along the time horizon. Our method, GATSBI, utilizes unsupervised object-centric scene representation learning to separate an active agent, static back-ground, and passive objects. GATSBI then models the interactions reflecting the causal relationships among decomposed entities and predicts physically plausible future states. Our model generalizes to a variety of environments where different types of robots and objects dynamically interact with each other. We show GATSBI achieves superior performance on scene decomposition and video prediction compared to its state-of-the-art counterparts.

Re-assembling multiple pots accurately from numerous 3D scanned fragments remains a challenging task to this date. Previous methods extract all potential matching pairs of pot sherds and considers them simultaneously to search for an optimal global pot configuration. In this work, we empirically show such global approach greatly suffers from false positive matches between sherds inflicted by indistinc-tive sharp fracture surfaces in pot fragments. To mitigate this problem, we take inspirations from the field of structure-from-motion (SfM), where many pipelines have matured in reconstructing a 3D scene from multiple images. Motivated by the success of the incremental approach in robust SfM, we present an efficient reassembly method for axially symmetric pots based on iterative registration of one sherd at a time. Our method goes beyond replicating incremental SfM and addresses indistinguishable false matches by embracing beam search to explore multitudes of registration possibilities. Additionally, we utilize multiple roots in each step to allow simultaneous reassembly of multiple pots. The proposed approach shows above 80% reassembly accuracy on a dataset of real 80 fragments mixed from 5 pots, pushing the state-of-the-art and paving the way towards the goal of large-scale pot reassembly. Our code and preprocessed data is available at https://github.com/SeongJong-Yoo/structure-from-sherds.

We introduce N-ImageNet, a large-scale dataset targeted for robust, fine-grained object recognition with event cameras. The dataset is collected using programmable hardware in which an event camera consistently moves around a monitor displaying images from ImageNet. N-ImageNet serves as a challenging benchmark for event-based object recognition, due to its large number of classes and samples. We empirically show that pretraining on N-ImageNet improves the performance of event-based classifiers and helps them learn with few labeled data. In addition, we present several variants of N-ImageNet to test the robustness of event-based classifiers under diverse camera trajectories and severe lighting conditions, and propose a novel event representation to alleviate the performance degradation. To the best of our knowledge, we are the first to quantitatively investigate the consequences caused by various environmental conditions on event-based object recognition algorithms. N-ImageNet and its variants are expected to guide practical implementations for deploying event-based object recognition algorithms in the real world.

© 2021 IEEEWe present PICCOLO, a simple and efficient algorithm for omnidirectional localization. Given a colored point cloud and a 360∘ panorama image of a scene, our objective is to recover the camera pose at which the panorama image is taken. Our pipeline works in an off-the-shelf manner with a single image given as a query and does not require any training of neural networks or collecting ground-truth poses of images. Instead, we match each point cloud color to the holistic view of the panorama image with gradient-descent optimization to find the camera pose. Our loss function, called sampling loss, is point cloud-centric, evaluated at the projected location of every point in the point cloud. In contrast, conventional photometric loss is image-centric, comparing colors at each pixel location. With a simple change in the compared entities, sampling loss effectively overcomes the severe visual distortion of omnidirectional images, and enjoys the global context of the 360∘ view to handle challenging scenarios for visual localization. PICCOLO outperforms existing omnidirectional localization algorithms in both accuracy and stability when evaluated in various environments.

Abstract We present the iterative medial axis transform (IMAT), an iterative descent method that constructs a medial axis transform (MAT) for a sparse, noisy, oriented point cloud sampled from an object's boundary. We first establish the equivalence between the traditional definition of the MAT of an object, i.e., the set of centres and corresponding radii of all balls maximally inscribed inside the object, with an alternative characterization matching the boundary enclosing the union of the balls with the object boundary. Based on this boundary equivalence characterization, a new MAT algorithm is proposed, in which an error function that reflects the difference between the two boundaries is minimized while restricting the number of balls to within some a priori specified upper limit. An iterative descent method with guaranteed local convergence is developed for the minimization that is also amenable to parallelization. Both quantitative and qualitative analyses of diverse 2D and 3D objects demonstrate the noise robustness, shape fidelity, and representation efficiency of the resulting MAT.

Scanning and acquiring a 3D indoor environment suffers from complex occlusions and misalignment errors. The reconstruction obtained from an RGB-D scanner contains holes in geometry and ghosting in texture. These are easily noticeable and cannot be considered as visually compelling VR content without further processing. On the other hand, the well-known Manhattan World priors successfully recreate relatively simple structures. In this article, we would like to push the limit of planar representation in indoor environments. Given an initial 3D reconstruction captured by an RGB-D sensor, we use planes not only to represent the environment geometrically but also to solve an inverse rendering problem considering texture and light. The complex process of shape inference and intrinsic imaging is greatly simplified with the help of detected planes and yet produces a realistic 3D indoor environment. The generated content can adequately represent the spatial arrangements for various AR/VR applications and can be readily composited with virtual objects possessing plausible lighting and texture.

Novel view synthesis is the task of synthesizing an image of an object at an arbitrary viewpoint given one or a few views of the object [1]. The output image of novel view synthesis exhibits a significant structural change from the input. Because of the large change, the skip connections or U-Net architecture, which can sustain the multi-level characteristics of the input images, cannot be directly utilized for the novel view synthesis [2]. In this paper, we investigate several variations of skip connection on two widely used novel view synthesis modules, pixel generation [1] and flow prediction [3]. For pixel generation, we find that the combination of the skip connections with flow-based hard attention is helpful. On the other hand, flow prediction enjoys marginal benefit from skip connections in deeper layers. Our pipeline suggests how to make use of skip connections on tasks that involve large geometric changes.

While potteries are common artifacts excavated in archaeological sites, the restoration process relies on manual cleaning and reassembly of shattered pieces. Since the number of possible 3D configurations is considerably large, the exhaustive manual trial may result in abrasion on fractured surfaces and even failure to find the correct matches. As a result, many recent works suggest virtual reassembly from 3D scans of the fragments. The problem is challenging in the view of the conventional 3D geometric analysis, as it is hard to extract reliable shape features from the thin break lines. We propose to optimize for the global configuration by combining geometric constraints with information from noisy shape features. Specifically, we enforce bijection and continuity of sequence of correspondences given estimates of corners and pair-wise matching scores between multiple break lines. We demonstrate that our pipeline greatly increases the accuracy of correspondences, resulting in stable restoration of 3D configurations from irregular and noisy evidence.

We present GROOT, a mobile volumetric video streaming system that delivers three-dimensional data to mobile devices for a fully immersive virtual and augmented reality experience. The system design for streaming volumetric videos should be fundamentally different from conventional 2D video streaming systems. First, the amount of data required to deliver the 3D volume is considerably larger than conventional videos with frames of 2D images, even compared to high-resolution 2D or 360° videos. Second, the 3D data representation, which encodes the surface of objects within the volume, is a sparse and unorganized data structure with varying scales, whereas a conventional video is composed of a sequence of images with the fixed-size 2D grid structure. GROOT is a streaming framework with a novel data structure that enables not only real-time transmission and decoding on mobile devices but also continuous on-demand user view adaptation. Specifically, we modify the conventional octree to introduce the independence of leaf nodes with minimal memory overhead, which enables parallel decoding of highly irregular 3D data. We also developed a suite of techniques to compress color information and filter out 3D points outside of a user's view, which efficiently minimizes the data size and decoding cost. Our extensive evaluation shows that GROOT achieves more stable but faster frame rates compared to any previous method to stream and visualize volumetric videos on mobile devices.

In this article, we present an automatic pose generation approach for robotic 3-D scanning systems designed for the reverse engineering of mechanical parts. 3-D scanning of an object is usually a highly complicated process that involves aligning multiple measurements from different viewpoints, fusing the merged data, and finally reconstructing the underlying surface. The reconstruction is often performed in isolation to the acquisition step. In contrast, the measured data often exhibit holes owing to limited viewpoints or noises in measurements, which can be observed after the postprocessing. Previous approaches either use the predefined scan paths without knowledge of the actual surface, or calculate the scan path based on the approximated smooth surface enclosing the incomplete scanned measurement. The need for an accurate 3-D scanning of mechanical parts often arises for reverse engineering purposes, as their shape often exhibits sharp corners or continuous shape of primitives (planes, cylinders, etc.), which the existing pipelines do not consider. We suggest a 3-D scan pipeline specifically designed to generate optimal scan paths of mechanical parts from incomplete data with limited fields of view. Our pipeline detects primitives from the scanned partial data, and utilizes the detection to approximate the underlying shape of the ground truth object; thus, it suggests the next-best-view of the scanner from the scarce measurements. In this regard, the suggested method is the first working pipeline specifically designed for scanning mechanical parts; furthermore, it outperforms previous approaches in terms of an efficient scan path generation.

We present RL-GAN-Net, where a reinforcement learning (RL) agent provides fast and robust control of a generative adversarial network (GAN). Our framework is applied to point cloud shape completion that converts noisy, partial point cloud data into a high-fidelity completed shape by controlling the GAN. While a GAN is unstable and hard to train, we circumvent the problem by (1) training the GAN on the latent space representation whose dimension is reduced compared to the raw point cloud input and (2) using an RL agent to find the correct input to the GAN to generate the latent space representation of the shape that best fits the current input of incomplete point cloud. The suggested pipeline robustly completes point cloud with large missing regions. To the best of our knowledge, this is the first attempt to train an RL agent to control the GAN, which effectively learns the highly nonlinear mapping from the input noise of the GAN to the latent space of point cloud. The RL agent replaces the need for complex optimization and consequently makes our technique real time. Additionally, we demonstrate that our pipelines can be used to enhance the classification accuracy of point cloud with missing data.

Reconstructing 3D models given a single-view 2D information is inherently an ill-posed problem and requires additional information such as shape prior or user input.We introduce a method to generate multiple 3D models of a particular category given corresponding photographs when the topological information is known. While there is a wide range of shapes for an object of a particular category, the basic topology usually remains constant.In consequence, the topological prior needs to be provided only once for each category and can be easily acquired by consulting an existing database of 3D models or by user input. The input of topological description is only connectivity information between parts; this is in contrast to previous approaches that have required users to interactively mark individual parts. Given the silhouette of an object and the topology, our system automatically finds a skeleton and generates a textured 3D model by jointly fitting multiple parts. The proposed method, therefore, opens the possibility of generating a large number of 3D models by consulting a massive number of photographs. We demonstrate examples of the topological prior and reconstructed 3D models using photos.

Abstract Advances in 3D acquisition devices provide unprecedented opportunities for quickly scanning indoor environments. Such raw scans, however, are often noisy, incomplete, and significantly corrupted, making semantic scene understanding difficult, if not impossible. Unfortunately, in most existing workflows, scan quality is assessed after the scanning stage is completed, making it cumbersome to correct for significant missing data by additional scanning. In this work, we present a guided real‐time scanning setup, wherein the incoming 3D data stream is continuously analyzed, and the data quality is automatically assessed. While the user is scanning an object, the proposed system discovers and highlights potential missing parts, thus guiding the operator (or an autonomous robot) as where to scan next. The proposed system assesses the quality and completeness of the 3D scan data by comparing to a large collection of commonly occurring indoor man‐made objects using an efficient, robust, and effective scan descriptor. We have tested the system on a large number of simulated and real setups, and found the guided interface to be effective even in cluttered and complex indoor environments.

We present a hand-held system for real-time, interactive acquisition of residential floor plans. The system integrates a commodity range camera, a micro-projector, and a button interface for user input and allows the user to freely move through a building to capture its important architectural elements. The system uses the Manhattan world assumption, which posits that wall layouts are rectilinear. This assumption allows generating floor plans in real time, enabling the operator to interactively guide the reconstruction process and to resolve structural ambiguities and errors during the acquisition. The interactive component aids users with no architectural training in acquiring wall layouts for their residences. We show a number of residential floor plans reconstructed with the system.

Large-scale acquisition of exterior urban environments is by now a well-established technology, supporting many applications in search, navigation, and commerce. The same is, however, not the case for indoor environments, where access is often restricted and the spaces are cluttered. Further, such environments typically contain a high density of repeated objects (e.g., tables, chairs, monitors, etc.) in regular or non-regular arrangements with significant pose variations and articulations. In this paper, we exploit the special structure of indoor environments to accelerate their 3D acquisition and recognition with a low-end handheld scanner. Our approach runs in two phases: (i) a learning phase wherein we acquire 3D models of frequently occurring objects and capture their variability modes from only a few scans, and (ii) a recognition phase wherein from a single scan of a new area, we identify previously seen objects but in different poses and locations at an average recognition time of 200ms/model. We evaluate the robustness and limits of the proposed recognition system using a range of synthetic and real world scans under challenging settings.