We introduce Any6D, a model-free framework for 6D object pose estimation that requires only a single RGB-D anchor image to estimate both the 6D pose and size of unknown objects in novel scenes. Unlike existing methods that rely on textured 3D models or multiple viewpoints, Any6D leverages a joint object alignment process to enhance 2D-3D alignment and metric scale estimation for improved pose accuracy. Our approach integrates a render-and-compare strategy to generate and refine pose hypotheses, enabling robust performance in scenarios with occlusions, non-overlapping views, diverse lighting conditions, and large cross-environment variations. We evaluate our method on five challenging datasets: REAL275, ToyotaLight, HO3D, YCBINEOAT, and LM-O, demonstrating its effectiveness in significantly outperforming state-of-the-art methods for novel object pose estimation. Project page: https://taeyeop.com/any6d

Distinguishing visually similar objects by their motion remains a critical challenge in computer vision. Although supervised trackers show promise, contemporary self-supervised trackers struggle when visual cues become ambiguous, limiting their scalability and generalization without extensive labeled data. We find that pre-trained video diffusion models inherently learn motion representations suitable for tracking without task-specific training. This ability arises because their denoising process isolates motion in early, high-noise stages, distinct from later appearance refinement. Capitalizing on this discovery, our self-supervised tracker significantly improves performance in distinguishing visually similar objects, an underexplored failure point for existing methods. Our method achieves up to a 6-point improvement over recent self-supervised approaches on established benchmarks and our newly introduced tests focused on tracking visually similar items. Visualizations confirm that these diffusion-derived motion representations enable robust tracking of even identical objects across challenging viewpoint changes and deformations. Project page: \small{\url{https://chenshuang-zhang.github.io/projects/ted}}.

Actual image super-resolution is an extremely challenging task due to complex degradations existing in the image. To solve this problem, two dominant methodologies have emerged: degradation-estimation-based Addressing actual image super-resolution remains a formidable challenge due to the intricate degradations present in images. Two primary methodologies have emerged: degradation-estimation-based and blind-based methods. The former often struggle to accurately estimate degradation, limiting their effectiveness on real low-resolution images. Conversely, blind-based methods rely on a single perceptual perspective, constraining their adaptability to diverse perceptual characteristics. In response to these challenges, we present MPF-Net, a novel super-resolution approach aimed at enhancing real-world image super-resolution tasks by enabling the model to learn multiple perceptual features from input images. Our method features a Multi-Perception Feature Extraction module (MPFE) designed to extract diverse perceptual details, complemented by Cross-Perception Blocks (CPB) facilitating the fusion of this information for efficient super-resolution reconstruction. Additionally, we introduce a contrastive regularization term (CR) to enhance the model’s learning by leveraging newly generated HR and LR images as positive and negative samples. Experimental results on challenging real-world SR datasets demonstrate the superiority of our approach over existing state-of-the-art methods, both qualitatively and quantitatively.

This paper proposes an algorithm for synthesizing novel views under few-shot setup. The main concept is to develop a stable surface regularization technique called Annealing Signed Distance Function (ASDF), which anneals the surface in a coarse-to-fine manner to accelerate convergence speed. We observe that the Eikonal loss – which is a widely known geometric regularization – requires dense training signal to shape different level-sets of SDF, leading to low-fidelity results under few-shot training. In contrast, the proposed surface regularization successfully reconstructs scenes and produce high-fidelity geometry with stable training. Our method is further accelerated by utilizing grid representation and monocular geometric priors. Finally, the proposed approach is up to 45 times faster than existing few-shot novel view synthesis methods, and it produces comparable results in the ScanNet dataset and NeRF-Real dataset.

Multi-target multi-camera tracking is a crucial task that involves identifying and tracking individuals over time using video streams from multiple cameras. This task has practical applications in various fields, such as visual surveillance, crowd behavior analysis, and anomaly detection. However, due to the difficulty and cost of collecting and labeling data, existing datasets for this task are either synthetically generated or artificially constructed within a controlled camera network setting, which limits their ability to model real-world dynamics and generalize to diverse camera configurations. To address this issue, we present MTMMC, a real-world, large-scale dataset that includes long video sequences captured by 16 multi-modal cameras in two different environments - campus and factory - across various time, weather, and season conditions. This dataset provides a challenging test-bed for studying multi-camera tracking under diverse real-world complexities and includes an additional input modality of spatially aligned and temporally synchronized RGB and thermal cameras, which enhances the accuracy of multi-camera tracking. MTMMC is a super-set of existing datasets, benefiting independent fields such as person detection, re-identification, and multiple object tracking. We provide baselines and new learning setups on this dataset and set the reference scores for future studies. The datasets, models, and test server will be made publicly available.

We establish rigorous benchmarks for visual perception robustness. Synthetic images such as ImageNet-C, ImageNet-9, and Stylized ImageNet provide specific type of evaluation over synthetic corruptions, backgrounds, and textures, yet those robustness benchmarks are restricted in specified variations and have low synthetic quality. In this work, we introduce generative model as a data source for synthesizing hard images that benchmark deep models' robustness. Leveraging diffusion models, we are able to generate images with more diversified backgrounds, textures, and materials than any prior work, where we term this benchmark as ImageNet-D. Experimental results show that ImageNet-D results in a significant accuracy drop to a range of vision models, from the standard ResNet visual classifier to the latest foundation models like CLIP and MiniGPT-4, significantly reducing their accuracy by up to 60%. Our work suggests that diffusion models can be an effective source to test vision models. The code and dataset are available at https://github.com/chenshuang-zhang/imagenet_d.

In this paper, we propose a new method to enhance compositional understanding in pretrained vision and language models (VLMs) without sacrificing performance in zero-shot multi-modal tasks.Traditional fine-tuning approaches often improve compositional reasoning at the cost of degrading multi-modal capabilities, primarily due to the use of global hard negative (HN) loss, which contrasts global representations of images and texts.This global HN loss pushes HN texts that are highly similar to the original ones, damaging the model's multi-modal representations.To overcome this limitation, we propose Fine-grained Selective Calibrated CLIP (FSC-CLIP), which integrates local hard negative loss and selective calibrated regularization.These innovations provide fine-grained negative supervision while preserving the model's representational integrity.Our extensive evaluations across diverse benchmarks for both compositionality and multi-modal tasks show that FSC-CLIP not only achieves compositionality on par with state-of-the-art models but also retains strong multi-modal capabilities.

RGB-thermal semantic segmentation is one potential solution to achieve reliable semantic scene understanding in adverse weather and lighting conditions. However, the previous studies mostly focus on designing a multi-modal fusion module without consideration of the nature of multi-modality inputs. Therefore, the networks easily become over-reliant on a single modality, making it difficult to learn complementary and meaningful representations for each modality. This paper proposes 1) a complementary random masking strategy of RGB-T images and 2) self-distillation loss between clean and masked input modalities. The proposed masking strategy prevents over-reliance on a single modality. It also improves the accuracy and robustness of the neural network by forcing the network to segment and classify objects even when one modality is partially available. Also, the proposed self-distillation loss encourages the network to extract complementary and meaningful representations from a single modality or complementary masked modalities. We achieve state-of-the-art performance over three RGB-T semantic segmentation benchmarks. Our source code is available at https://github.com/UkcheolShin/CRM_RGBTSeg.

Existing video compression (VC) methods primarily aim to reduce the spatial and temporal redundancies between consecutive frames in a video while preserving its quality. In this regard, previous works have achieved remarkable results on videos acquired under specific settings such as instant (known) exposure time and shutter speed which often result in sharp videos. However, when these methods are evaluated on videos captured under different temporal priors, which lead to degradations like motion blur and low frame rate, they fail to maintain the quality of the contents. In this work, we tackle the VC problem in a general scenario where a given video can be blurry due to predefined camera settings or dynamics in the scene. By exploiting the natural trade-off between visual enhancement and data compression, we formulate VC as a min-max optimization problem and propose an effective framework and training strategy to tackle the problem. Extensive experimental results on several benchmark datasets confirm the effectiveness of our method compared to several state-of-the-art VC approaches.

Video Panoptic Segmentation (VPS) aims to achieve comprehensive pixel-level scene understanding by segmenting all pixels and associating objects in a video. Current solutions can be categorized into online and near-online approaches. Evolving over the time, each category has its own specialized designs, making it nontrivial to adapt models between different categories. To alleviate the discrepancy, in this work, we propose a unified approach for online and near-online VPS. The meta architecture of the proposed Video-kMaX consists of two components: within-clip segmenter (for clip-level segmentation) and cross-clip associater (for association beyond clips). We propose clip-kMaX (clip k-means mask transformer) and LA-MB (location-aware memory buffer) to instantiate the segmenter and associater, respectively. Our general formulation includes the online scenario as a special case by adopting clip length of one. Without bells and whistles, Video-kMaX sets a new state-of-the-art on KITTI-STEP and VIPSeg for video panoptic segmentation Code and models are available at this link.

Mixup provides interpolated training samples and allows the model to obtain smoother decision boundaries for better generalization. The idea can be naturally applied to the domain adaptation task, where we can mix the source and target samples to obtain domain-mixed samples for better adaptation. However, the extension of the idea from classification to segmentation (i.e., structured output) is nontrivial. This paper systematically studies the impact of mixup under the domain adaptive semantic segmentation task and presents a simple yet effective mixup strategy called Bidirectional Domain Mixup (BDM). In specific, we achieve domain mixup in two-step: cut and paste. Given the warm-up model trained from any adaptation techniques, we forward the source and target samples and perform a simple threshold-based cut out of the unconfident regions (cut). After then, we fill-in the dropped regions with the other domain region patches (paste). In doing so, we jointly consider class distribution, spatial structure, and pseudo label confidence. Based on our analysis, we found that BDM leaves domain transferable regions by cutting, balances the dataset-level class distribution while preserving natural scene context by pasting. We coupled our proposal with various state-of-the-art adaptation models and observe significant improvement consistently. We also provide extensive ablation experiments to empirically verify our main components of the framework. Visit our project page with the code at https://sites.google.com/view/bidirectional-domain-mixup

The task of Visual Question Answering (VQA) is known to be plagued by the issue of VQA models exploiting biases within the dataset to make its final prediction. Various previous ensemble based debiasing methods have been proposed where an additional model is purposefully trained to be biased in order to train a robust target model. However, these methods compute the bias for a model simply from the label statistics of the training data or from single modal branches. In this work, in order to better learn the bias a target VQA model suffers from, we propose a generative method to train the bias model directly from the target model, called GenB. In particular, GenB employs a generative network to learn the bias in the target model through a combination of the adversarial objective and knowledge distillation. We then debias our target model with GenB as a bias model, and show through extensive experiments the effects of our method on various VQA bias datasets including VQA-CP2, VQA-CP1, GQA-OOD, and VQA-CE, and show state-of-the-art results with the LXMERT architecture on VQA-CP2.

Test-time adaptation methods have been gaining attention recently as a practical solution for addressing source-to-target domain gaps by gradually updating the model without requiring labels on the target data. In this paper, we propose a method of test-time adaptation for category-level object pose estimation called TTA-COPE. We design a pose ensemble approach with a self-training loss using pose-aware confidence. Unlike previous unsupervised domain adaptation methods for category-level object pose estimation, our approach processes the test data in a sequential, online manner, and it does not require access to the source domain at runtime. Extensive experimental results demonstrate that the proposed pose ensemble and the self-training loss improve category-level object pose performance during test time under both semi-supervised and unsupervised settings.

In this paper, we propose a single image scale estimation method based on a novel scale field representation. A scale field defines the local pixel-to-metric conversion ratio along the gravity direction on all the ground pixels. This representation resolves the ambiguity in camera parameters, allowing us to use a simple yet effective way to collect scale annotations on arbitrary images from human annotators. By training our model on calibrated panoramic image data and the in-the-wild human annotated data, our single image scene scale estimation network generates robust scale field on a variety of image, which can be utilized in various 3D understanding and scale-aware image editing applications.

Mask-guided matting has shown great practicality compared to traditional trimap-based methods. The mask-guided approach takes an easily-obtainable coarse mask as guidance and produces an accurate alpha matte. To extend the success toward practical usage, we tackle mask-guided matting in the wild, which covers a wide range of categories in their complex context robustly. To this end, we propose a simple yet effective learning framework based on two core insights: 1) learning a generalized matting model that can better understand the given mask guidance and 2) leveraging weak supervision datasets (e.g., instance segmentation dataset) to alleviate the limited diversity and scale of existing matting datasets. Extensive experimental results on multiple benchmarks, consisting of a newly proposed synthetic benchmark (Composition-Wild) and existing natural datasets, demonstrate the superiority of the proposed method. Moreover, we provide appealing results on new practical applications (e.g., panoptic matting and mask-guided video matting), showing the great generality and potential of our model.

Robust and accurate geometric understanding against adverse weather conditions is one top prioritized conditions to achieve a high-level autonomy of self-driving cars. However, autonomous driving algorithms relying on the visible spectrum band are easily impacted by weather and lighting conditions. A long-wave infrared camera, also known as a thermal imaging camera, is a potential rescue to achieve high-level robustness. However, the missing necessities are the well-established large-scale dataset and public benchmark results. To this end, in this paper, we first built a large-scale Multi-Spectral Stereo (MS2) dataset, including stereo RGB, stereo NIR, stereo thermal, and stereo LiDAR data along with GNSS/IMU information. The collected dataset provides about 195K synchronized data pairs taken from city, residential, road, campus, and suburban areas in the morning, daytime, and nighttime under clear-sky, cloudy, and rainy conditions. Secondly, we conduct an exhaustive validation process of monocular and stereo depth estimation algorithms designed on visible spectrum bands to benchmark their performance in the thermal image domain. Lastly, we propose a unified depth network that effectively bridges monocular depth and stereo depth tasks from a conditional random field approach perspective. Our dataset and source code are available at https://github.com/UkcheolShin/MS2-MultiSpectralStereoDataset.

This paper presents a simple yet effective approach that improves continual test-time adaptation (TTA) in a memory-efficient manner. TTA may primarily be conducted on edge devices with limited memory, so reducing memory is crucial but has been overlooked in previous TTA studies. In addition, long-term adaptation often leads to catastrophic forgetting and error accumulation, which hinders applying TTA in real-world deployments. Our approach consists of two components to address these issues. First, we present lightweight meta networks that can adapt the frozen original networks to the target domain. This novel architecture minimizes memory consumption by decreasing the size of intermediate activations required for backpropagation. Second, our novel self-distilled regularization controls the output of the meta networks not to deviate significantly from the output of the frozen original networks, thereby preserving well-trained knowledge from the source domain. Without additional memory, this regularization prevents error accumulation and catastrophic forgetting, resulting in stable performance even in long-term test-time adaptation. We demonstrate that our simple yet effective strategy outperforms other state-of-the-art methods on various benchmarks for image classification and semantic segmentation tasks. Notably, our proposed method with ResNet-50 and WideResNet-40 takes 86% and 80% less memory than the recent state-of-the-art method, CoTTA.

Driven by improved architectures and better representation learning frameworks, the field of visual recognition has enjoyed rapid modernization and performance boost in the early 2020s. For example, modern ConvNets, represented by ConvNeXt [33], have demonstrated strong performance in various scenarios. While these models were originally designed for supervised learning with ImageNet labels, they can also potentially benefit from self-supervised learning techniques such as masked autoencoders (MAE) [14]. However, we found that simply combining these two approaches leads to subpar performance. In this paper, we propose a fully convolutional masked autoencoder framework and a new Global Response Normalization (GRN) layer that can be added to the ConvNeXt architecture to enhance inter-channel feature competition. This co-design of self-supervised learning techniques and architectural improvement results in a new model family called ConvNeXt V2, which significantly improves the performance of pure ConvNets on various recognition benchmarks, including ImageNet classification, COCO detection, and ADE20K segmentation. We also provide pre-trained ConvNeXt V2 models of various sizes, ranging from an efficient 3.7M-parameter Atto model with 76.7% top-1 accuracy on ImageNet, to a 650M Huge model that achieves a state-of-the-art 88.9% accuracy using only public training data.

The goal of this work is to develop self-sufficient framework for Continuous Sign Language Recognition (CSLR) that addresses key issues of sign language recognition. These include the need for complex multi-scale features such as hands, face, and mouth for understanding, and absence of frame-level annotations. To this end, we propose (1) Divide and Focus Convolution (DFConv) which extracts both manual and non-manual features without the need for additional networks or annotations, and (2) Dense Pseudo-Label Refinement (DPLR) which propagates non-spiky frame-level pseudo-labels by combining the ground truth gloss sequence labels with the predicted sequence. We demonstrate that our model achieves state-of-the-art performance among RGB-based methods on large-scale CSLR benchmarks, PHOENIX-2014 and PHOENIX-2014-T, while showing comparable results with better efficiency when compared to other approaches that use multi-modality or extra annotations.

The objective of this work is to explore the learning of visually grounded speech models (VGS) from multilingual perspective. Bilingual VGS models are generally trained with an equal number of spoken captions from both languages. However, in reality, there can be an imbalance among the languages for the available spoken captions. Our key contribution in this work is to leverage the power of a high-resource language in a bilingual visually grounded speech model to improve the performance of a low-resource language. We introduce two methods to distill the knowledge of high-resource language into low-resource languages: (1) incorporating a strong pre-trained high-resource language encoder and (2) using semantically similar spoken captions. Our experiments show that combining these two approaches effectively enables the low-resource language to surpass the performances of monolingual and bilingual counterparts for cross-modal retrieval tasks.

Learning computer vision models from (and for) movies has a long-standing history. While great progress has been attained, there is still a need for a pretrained multimodal model that can perform well in the ever-growing set of movie understanding tasks the community has been establishing. In this work, we introduce Long-range Multimodal Pretraining, a strategy, and a model that leverages movie data to train transferable multimodal and cross-modal encoders. Our key idea is to learn from all modalities in a movie by observing and extracting relationships over a long-range. After pretraining, we run ablation studies on the LVU benchmark and validate our modeling choices and the importance of learning from long-range time spans. Our model achieves state-of-the-art on several LVU tasks while being much more data efficient than previous works. Finally, we evaluate our model’s transferability by setting a new state-of-the-art in five different benchmarks.

We propose an algorithm, 4DRegSDF, for the spacetime surface regularization to improve the fidelity of neural rendering and reconstruction in dynamic scenes. The key idea is to impose local rigidity on the deformable Signed Distance Function (SDF) for temporal coherency. Our approach works by (1) sampling points on the deformed surface by taking gradient steps toward the steepest direction along SDF, (2) extracting differential surface geometry, such as tangent plane or curvature, at each sample, and (3) adjusting the local rigidity at different timestamps. This enables our dynamic surface regularization to align 4D spacetime geometry via 3D canonical space more accurately. Experiments demonstrate that our 4DRegSDF achieves state-of-the-art performance in both reconstruction and rendering quality over synthetic and real-world datasets. https://4dregsdf.github.io/

Our MATE is the first Test-Time-Training (TTT) method designed for 3D data, which makes deep networks trained for point cloud classification robust to distribution shifts occurring in test data. Like existing TTT methods from the 2D image domain, MATE also leverages test data for adaptation. Its test-time objective is that of a Masked Autoencoder: a large portion of each test point cloud is removed before it is fed to the network, tasked with reconstructing the full point cloud. Once the network is updated, it is used to classify the point cloud. We test MATE on several 3D object classification datasets and show that it significantly improves robustness of deep networks to several types of corruptions commonly occurring in 3D point clouds. We show that MATE is very efficient in terms of the fraction of points it needs for the adaptation. It can effectively adapt given as few as 5% of tokens of each test sample, making it extremely lightweight. Our experiments show that MATE also achieves competitive performance by adapting sparsely on the test data, which further reduces its computational overhead, making it ideal for real-time applications.

Diffusion probabilistic models (DPM) have been widely adopted in image-to-image translation to generate high-quality images. Prior attempts at applying the DPM to image super-resolution (SR) have shown that iteratively refining a pure Gaussian noise with a conditional image using a U-Net trained on denoising at various-level noises can help obtain a satisfied high-resolution image for the low-resolution one. To further improve the performance and simplify current DPM-based super-resolution methods, we propose a simple but non-trivial DPM-based super-resolution post-process framework, i.e., cDPMSR. After applying a pre-trained SR model on the to-be-test LR image to provide the conditional input, we adapt the standard DPM to conduct conditional image generation and perform super-resolution through a deterministic iterative denoising process. Our method surpasses prior attempts on both qualitative and quantitative results and can generate more photo-realistic counterparts for the low-resolution images with various benchmark datasets including Set5, Set14, Urban100, BSD100, and Manga109. Code will be published after accepted.

With the increasing popularity of masked autoencoders, self-supervised learning (SSL) in vision undertakes a similar trajectory as in NLP. Specifically, generative pretext tasks with the masked prediction have become a de facto standard SSL practice in NLP (e.g., BERT). By contrast, early attempts at generative methods in vision have been outperformed by their discriminative counterparts (like contrastive learning). However, the success of masked image modeling has revived the autoencoder-based visual pretraining method. As a milestone to bridge the gap with BERT in NLP, masked autoencoder in vision has attracted unprecedented attention. This work conducts a survey on masked autoencoders for visual SSL.

Transferable targeted adversarial attack against deep image classifiers has remained an open issue. Depending on the space to optimize the loss, the existing methods can be divided into two categories: (a) feature space attack and (b) output space attack. The feature space attack outperforms output space one by a large margin but at the cost of requiring the training of layer-wise auxiliary classifiers for each corresponding target class together with the greedy search for the optimal layers. In this work, we revisit the method of output space attack and improve it from two perspectives. First, we identify over-fitting as one major factor that hinders transferability, for which we propose to augment the network input and/or feature layers with noise. Second, we propose a new cross-entropy loss with two ends: one for pushing the sample far from the source class, i.e. ground-truth class, and the other for pulling it close to the target class. We demonstrate that simple techniques are sufficient enough for achieving very competitive performance.

Universal Domain Adaptation aims to transfer the knowledge between the datasets by handling two shifts: domain-shift and category-shift. The main challenge is correctly distinguishing the unknown target samples while adapting the distribution of known class knowledge from source to target. Most existing methods approach this problem by first training the target adapted known classifier and then relying on the single threshold to distinguish unknown target samples. However, this simple threshold-based approach prevents the model from considering the underlying complexities existing between the known and unknown samples in the high-dimensional feature space. In this paper, we propose a new approach in which we use two sets of feature points, namely dual Classifiers for Prototypes and Reciprocals (CPR). Our key idea is to associate each prototype with corresponding known class features while pushing the reciprocals apart from these prototypes to locate them in the potential unknown feature space. The target samples are then classified as unknown if they fall near any reciprocals at test time. To successfully train our framework, we collect the partial, confident target samples that are classified as known or unknown through on our proposed multi-criteria selection. We then additionally apply the entropy loss regularization to them. For further adaptation, we also apply standard consistency regularization that matches the predictions of two different views of the input to make more compact target feature space. We evaluate our proposal, CPR, on three standard benchmarks and achieve comparable or new state-of-the-art results. We also provide extensive ablation experiments to verify our main design choices in our framework.

To understand our surrounding world, our brain is continuously inundated with multisensory information and their complex interactions coming from the outside world at any given moment. While processing this information might seem effortless for human brains, it is challenging to build a machine that can perform similar tasks since complex interactions cannot be dealt with a single type of integration but require more sophisticated approaches. In this paper, we propose a new simple method to address the multisensory integration in video understanding. Unlike previous works where a single fusion type is used, we design a multi-head model with individual event-specific layers to deal with different audio-visual relationships, enabling different ways of audio-visual fusion. Experimental results show that our event-specific layers can discover unique properties of the audio-visual relationships in the videos, e.g., semantically matched moments, and rhythmic events. Moreover, although our network is trained with single labels, our multi-head design can inherently output additional semantically meaningful multi-labels for a video. As an application, we demonstrate that our proposed method can expose the extent of event-characteristics of popular benchmark datasets.

Recently, thermal image based 3D understanding is gradually attracting attention for an illumination condition agnostic machine vision. However, the difficulty of the thermal image lies in insufficient training supervision due to its low-contrast and texturesless properties. Also, introducing additional modality requires further constraints such as complicated multi-sensor calibration and synchronized data acquisition. To leverage additional modality information without such constraints, we propose a novel training framework that consists of self-supervised learning of unpaired multi-spectral images and feature-level adversarial adaptation. In the training stage, we utilize unpaired RGB/thermal video and partially shared network architecture consisting of modality-specific feature extractors and modality-independent decoder. Through the shared network design, the depth decoder can leverage the self-supervised signal of the unpaired RGB images. Feature-level adversarial adaptation minimizes the gap between RGB and thermal features and eventually makes the thermal encoder extract representative and informative features. Based on the proposed method, the trained depth network shows outperformed results than previous state-of-the-art methods.

With the increasing social demands of disaster response, methods of visual observation for rescue and safety have become increasingly important. However, because of the shortage of datasets for disaster scenarios, there has been little progress in computer vision and robotics in this field. With this in mind, we present the first large-scale synthetic dataset of egocentric viewpoints for disaster scenarios. We simulate pre- and post-disaster cases with drastic changes in appearance, such as buildings on fire and earthquakes. The dataset consists of more than 300K high-resolution stereo image pairs, all annotated with ground-truth data for the semantic label, depth in metric scale, optical flow with sub-pixel precision, and surface normal as well as their corresponding camera poses. To create realistic disaster scenes, we manually augment the effects with 3D models using physically-based graphics tools. We train various state-of-the-art methods to perform computer vision tasks using our dataset, evaluate how well these methods recognize the disaster situations, and produce reliable results of virtual scenes as well as real-world images. We also present a convolutional neural network-based egocentric localization method that is robust to drastic appearance changes, such as the texture changes in a fire, and layout changes from a collapse. To address these key challenges, we propose a new model that learns a shape-based representation by training on stylized images, and incorporate the dominant planes of query images as approximate scene coordinates. We evaluate the proposed method using various scenes including a simulated disaster dataset to demonstrate the effectiveness of our method when confronted with significant changes in scene layout. Experimental results show that our method provides reliable camera pose predictions despite vastly changed conditions.

Thanks to Earth-level Street View images from Google Maps, a visual image geo-localization can estimate the coarse location of a query image with a visual place recognition process. However, this can get very challenging when non-static objects change with time, severely degrading image retrieval accuracy. We address the problem of city-scale visual place recognition in complex urban environments crowded with non-static clutters. To this end, we first analyze what clutters degrade similarity matching between the query and database images. Second, we design a self-supervised trainable de-attention module that prevents the network from focusing on non-static objects in an input image. In addition, we propose a novel triplet marginal loss called sharpened triplet marginal loss to make feature descriptors more discriminative. Lastly, due to the lack of geo-tagged public datasets with a high density of non-static objects, we propose a clutter augmentation method to evaluate our approach. The experimental results show that our model has notably improved over the existing attention methods in geo-localization tasks on the public benchmark datasets and on their augmented versions with high population and traffic. Our code is available at https://github.com/ccsmm78/deattention_with_stml_for_vpr.

Recent state-of-the-art active learning methods have mostly leveraged generative adversarial networks (GANs) for sample acquisition; however, GAN is usually known to suffer from instability and sensitivity to hyperparameters. In contrast to these methods, in this article, we propose a novel active learning framework that we call Maximum Classifier Discrepancy for Active Learning (MCDAL) that takes the prediction discrepancies between multiple classifiers. In particular, we utilize two auxiliary classification layers that learn tighter decision boundaries by maximizing the discrepancies among them. Intuitively, the discrepancies in the auxiliary classification layers' predictions indicate the uncertainty in the prediction. In this regard, we propose a novel method to leverage the classifier discrepancies for the acquisition function for active learning. We also provide an interpretation of our idea in relation to existing GAN-based active learning methods and domain adaptation frameworks. Moreover, we empirically demonstrate the utility of our approach where the performance of our approach exceeds the state-of-the-art methods on several image classification and semantic segmentation datasets in active learning setups.

Video frame interpolation (VFI) works generally predict intermediate frame(s) by first estimating the motion between inputs and then warping the inputs to the target time with the estimated motion. This approach, however, is not optimal when the temporal distance between the input sequence increases as existing motion estimation modules cannot effectively handle large motions. Hence, VFI works perform well for small frame gaps and perform poorly as the frame gap increases. In this work, we propose a novel framework to address this problem. We argue that when there is a large gap between inputs, instead of estimating imprecise motion that will eventually lead to inaccurate interpolation, we can safely propagate from one side of the input up to a reliable time frame using the other input as a reference. Then, the rest of the intermediate frames can be interpolated using standard approaches as the temporal gap is now narrowed. To this end, we propose a propagation network (PNet) by extending the classic feature-level forecasting with a novel motion-to-feature approach. To be thorough, we adopt a simple interpolation model along with PNet as our full model and design a simple procedure to train the full model in an end-to-end manner. Experimental results on several benchmark datasets confirm the effectiveness of our method for long-term VFI compared to state-of-the-art approaches.

We present TubeFormer-DeepLab, the first attempt to tackle multiple core video segmentation tasks in a unified manner. Different video segmentation tasks (e.g., video semantic/instance/panoptic segmentation) are usually considered as distinct problems. State-of-the-art models adopted in the separate communities have diverged, and radically different approaches dominate in each task. By contrast, we make a crucial observation that video segmentation tasks could be generally formulated as the problem of assigning different predicted labels to video tubes (where a tube is obtained by linking segmentation masks along the time axis) and the labels may encode different values depending on the target task. The observation motivates us to develop TubeFormer-DeepLab, a simple and effective video mask transformer model that is widely applicable to multiple video segmentation tasks. TubeFormer-DeepLab directly predicts video tubes with task-specific labels (either pure semantic categories, or both semantic categories and instance identities), which not only significantly simplifies video segmentation models, but also advances state-of-the-art results on multiple video segmentation benchmarks.

Learning to estimate object pose often requires ground-truth (GT) labels, such as CAD model and absolute-scale object pose, which is expensive and laborious to obtain in the real world. To tackle this problem, we propose an unsupervised domain adaptation (UDA) for category-level object pose estimation, called UDA-COPE. Inspired by recent multi-modal UDA techniques, the proposed method exploits a teacher-student self-supervised learning scheme to train a pose estimation network without using target domain pose labels. We also introduce a bidirectional filtering method between the predicted normalized object coordinate space (NOCS) map and observed point cloud, to not only make our teacher network more robust to the target domain but also to provide more reliable pseudo labels for the student network training. Extensive experimental results demonstrate the effectiveness of our proposed method both quantitatively and qualitatively. Notably, without leveraging target-domain GT labels, our proposed method achieved comparable or sometimes superior performance to existing methods that depend on the GT labels.

The capability of the traditional semi-supervised learning (SSL) methods is far from real-world application due to severely biased pseudo-labels caused by (1) class imbalance and (2) class distribution mismatch between labeled and unlabeled data. This paper addresses such a relatively under-explored problem. First, we propose a general pseudo-labeling framework that class-adaptively blends the semantic pseudo-label from a similarity-based classifier to the linear one from the linear classifier, after making the observation that both types of pseudo-labels have complementary properties in terms of bias. We further introduce a novel semantic alignment loss to establish balanced feature representation to reduce the biased predictions from the classifier. We term the whole framework as Distribution-Aware Semantics-Oriented (DASO) Pseudo-label. We conduct extensive experiments in a wide range of imbalanced benchmarks: CIFAR10/100-LT, STL10-LT, and large-scale long-tailed Semi-Aves with open-set class, and demonstrate that, the proposed DASO framework reliably improves SSL learners with unlabeled data especially when both (1) class imbalance and (2) distribution mismatch dominate.

Recently, memory-based approaches show promising results on semi-supervised video object segmentation. These methods predict object masks frame-by-frame with the help of frequently updated memory of the previous mask. Different from this per-frame inference, we investigate an alternative perspective by treating video object segmentation as clip-wise mask propagation. In this per-clip inference scheme, we update the memory with an interval and simul-taneously process a set of consecutive frames (i.e. clip) between the memory updates. The scheme provides two potential benefits: accuracy gain by clip-level optimization and efficiency gain by parallel computation of multiple frames. To this end, we propose a new method tailored for the perclip inference. Specifically, we first introduce a clip-wise operation to refine the features based on intra-clip correlation. In addition, we employ a progressive matching mechanism for efficient information-passing within a clip. With the synergy of two modules and a newly proposed perclip based training, our network achieves state-of-the-art performance on Youtube-VOS 2018/2019 val (84.6% and 84.6%) and DAVIS 2016/2017 val (91.9% and 86.1%). Fur-thermore, our model shows a great speed-accuracy trade-off with varying memory update intervals, which leads to huge flexibility.

Test-time adaptation approaches have recently emerged as a practical solution for handling domain shift without access to the source domain data. In this paper, we propose and explore a new multi-modal extension of test-time adaptation for 3D semantic segmentation. We find that, directly applying existing methods usually results in performance instability at test time, because multi-modal input is not considered jointly. To design a framework that can take full advantage of multi-modality, where each modality provides regularized self-supervisory signals to other modalities, we propose two complementary modules within and across the modalities. First, Intra-modal Pseudo-label Generation (Intra-PG) is introduced to obtain reliable pseudo labels within each modality by aggregating information from two models that are both pre-trained on source data but updated with target data at different paces. Second, Inter-modal Pseudo-label Refinement (Inter-PR) adaptively selects more reliable pseudo labels from different modalities based on a proposed consistency scheme. Experiments demonstrate that our regularized pseudo labels produce stable self-learning signals in numerous multi-modal test-time adaptation scenarios for 3D semantic segmentation. Visit our project website at https://www.nec-labs.com/~mas/MM-TTA

Existing works have identified the limitation of top-1 attack success rate (ASR) as a metric to evaluate the attack strength but exclusively investigated it in the white-box setting, while our work extends it to a more practical black-box setting: transferable attack. It is widely reported that stronger I-FGSM transfers worse than simple FGSM, leading to a popular belief that transferability is at odds with the white-box attack strength. Our work challenges this belief with empirical finding that stronger attack actually transfers better for the general top-k ASR indicated by the interest class rank (ICR) after attack. For increasing the attack strength, with an intuitive analysis on the logit gradient from the geometric perspective, we identify that the weakness of the commonly used losses lie in prioritizing the speed to fool the network instead of maximizing its strength. To this end, we propose a new normalized CE loss that guides the logit to be updated in the direction of implicitly maximizing its rank distance from the ground-truth class. Extensive results in various settings have verified that our proposed new loss is simple yet effective for top-k attack. Code is available at: https://bit.ly/3uCiomP

Contrastive learning (CL) is widely known to require many negative samples, 65536 in MoCo for instance, for which the performance of a dictionary-free framework is often inferior because the negative sample size (NSS) is limited by its mini-batch size (MBS). To decouple the NSS from the MBS, a dynamic dictionary has been adopted in a large volume of CL frameworks, among which arguably the most popular one is MoCo family. In essence, MoCo adopts a momentum-based queue dictionary, for which we perform a fine-grained analysis of its size and consistency. We point out that InfoNCE loss used in MoCo implicitly attract anchors to their corresponding positive sample with various strength of penalties and identify such inter-anchor hardness-awareness property as a major reason for the necessity of a large dictionary. Our findings motivate us to simplify MoCo v2 via the removal of its dictionary as well as momentum. Based on an InfoNCE with the proposed dual temperature, our simplified frameworks, Sim-MoCo and SimCo, outperform MoCo v2 by a visible margin. Moreover, our work bridges the gap between CL and non-CL frameworks, contributing to a more unified under-standing of these two mainstream frameworks in SSL. Code is available at: https://bit.ly/3LkQbaT.

The objective of this work is to localize the sound sources in visual scenes. Existing audio-visual works employ contrastive learning by assigning corresponding audio-visual pairs from the same source as positives while randomly mismatched pairs as negatives. However, these negative pairs may contain semantically matched audio-visual information. Thus, these semantically correlated pairs, "hard positives", are mistakenly grouped as negatives. Our key contribution is showing that hard positives can give similar response maps to the corresponding pairs. Our approach incorporates these hard positives by adding their response maps into a contrastive learning objective directly. We demonstrate the effectiveness of our approach on VGG-SS and SoundNet-Flickr test sets, showing favorable performance to the state-of-the-art methods.

In this paper, we propose a multi-objective camera ISP framework that utilizes Deep Reinforcement Learning (DRL) and camera ISP toolbox that consist of network-based and conventional ISP tools. The proposed DRL-based camera ISP framework iteratively selects a proper tool from the toolbox and applies it to the image to maximize a given vision task-specific reward function. For this purpose, we implement total 51 ISP tools that include exposure correction, color-and-tone correction, white balance, sharpening, denoising, and the others. We also propose an efficient DRL network architecture that can extract the various aspects of an image and make a rigid mapping relationship between images and a large number of actions. Our proposed DRL-based ISP framework effectively improves the image quality according to each vision task such as RAW-to-RGB image restoration, 2D object detection, and monocular depth estimation.

In this paper, we tackle sound localization as a natural outcome of the audio-visual video classification problem. Differently from the existing sound localization approaches, we do not use any explicit sub-modules or training mechanisms but use simple cross-modal attention on top of the representations learned by a classification loss. Our key contribution is to show that a simple audio-visual classification model has the ability to localize sound sources accurately and to give on par performance with state-of-the-art methods by proving that indeed "less is more". Furthermore, we propose potential applications that can be built based on our model. First, we introduce informative moment selection to enhance the localization task learning in the existing approaches compare to mid-frame usage. Then, we introduce a pseudo bounding box generation procedure that can significantly boost the performance of the existing methods in semi-supervised settings or be used for large-scale automatic annotation with minimal effort from any video dataset.

We propose a new linear RGB-D simultaneous localization and mapping (SLAM) formulation by utilizing planar features of the structured environments. The key idea is to understand a given structured scene and exploit its structural regularities such as the Manhattan world. This understanding allows us to decouple the camera rotation by tracking structural regularities, which makes SLAM problems free from being highly nonlinear. Additionally, it provides a simple yet effective cue for representing planar features, which leads to a linear SLAM formulation. Given an accurate camera rotation, we jointly estimate the camera translation and planar landmarks in the global planar map using a linear Kalman filter. Our linear SLAM method, called L-SLAM, can understand not only the Manhattan world but the more general scenario of the Atlanta world, which consists of a vertical direction and a set of horizontal directions orthogonal to the vertical direction. To this end, we introduce a novel tracking-by-detection scheme that infers the underlying scene structure by Atlanta representation. With efficient Atlanta representation, we formulate a unified linear SLAM framework for structured environments. We evaluate L-SLAM on a synthetic dataset and RGB-D benchmarks, demonstrating comparable performance to other state-of-the-art SLAM methods without using expensive nonlinear optimization. We assess the accuracy of L-SLAM on a practical application of augmented reality.

Taking selfies has become one of the major photographic trends of our time. In this study, we focus on the selfie stick, on which a camera is mounted to take selfies. We observe that a camera on a selfie stick typically travels through a particular type of trajectory around a sphere. Based on this finding, we propose a robust, efficient, and optimal estimation method for relative camera pose between two images captured by a camera mounted on a selfie stick. We exploit the special geometric structure of camera motion constrained by a selfie stick and define this motion as spherical joint motion. Utilizing a novel parametrization and calibration scheme, we demonstrate that the pose estimation problem can be reduced to a 3-degrees of freedom (DoF) search problem, instead of a generic 6-DoF problem. This facilitates the derivation of an efficient branch-and-bound optimization method that guarantees a global optimal solution, even in the presence of outliers. Furthermore, as a simplified case of spherical joint motion, we introduce selfie motion, which has a fewer number of DoF than spherical joint motion. We validate the performance and guaranteed optimality of our method on both synthetic and real-world data. Additionally, we demonstrate the applicability of the proposed method for two applications: refocusing and stylization.

We introduce dense relational captioning, a novel image captioning task which aims to generate multiple captions with respect to relational information between objects in a visual scene. Relational captioning provides explicit descriptions for each relationship between object combinations. This framework is advantageous in both diversity and amount of information, leading to a comprehensive image understanding based on relationships, e.g., relational proposal generation. For relational understanding between objects, the part-of-speech (POS; i.e., subject-object-predicate categories) can be a valuable prior information to guide the causal sequence of words in a caption. We enforce our framework to learn not only to generate captions but also to understand the POS of each word. To this end, we propose the multi-task triple-stream network (MTTSNet) which consists of three recurrent units responsible for each POS which is trained by jointly predicting the correct captions and POS for each word. In addition, we found that the performance of MTTSNet can be improved by modulating the object embeddings with an explicit relational module. We demonstrate that our proposed model can generate more diverse and richer captions, via extensive experimental analysis on large scale datasets and several metrics. Then, we present applications of our framework to holistic image captioning, scene graph generation, and retrieval tasks.

At present, most mainstream algorithms for single image super-resolution (SISR) assume the image degradation process as an ideal degradation process (e.g. bicubic downscaling), which violates the actual degeneration conditions. In real-world image capturing, objects often move in a dynamic environment, and camera shake also often occurs, which results in serious blurs. Our work focuses on the task of image super-resolution with heavy motion blur, for which we adopt a network with two branches: one branch for image deblurring and the other one for super-resolution. Since the features obtained by the deblurring are rich in details, we apply their features as supplementary information to the super-resolution branch. Based on the adopted dual-branch framework, our major technical novelties lie in two novel modules: Multi-Scale Feature Fusion (MSFF1) module which fuses features of different scale from the deblurring branch to get local and global information, and Multi-Stage Feature Fusion (MSFF2) module which further filters useful information with attention. We evaluate the proposed method under various blur scenarios on the benchmark datasets, demonstrating competitive performance against existing methods.

A holistic understanding of dynamic scenes is of fundamental importance in real-world computer vision problems such as autonomous driving, augmented reality and spatio-temporal reasoning. In this paper, we propose a new computer vision benchmark: Video Panoptic Segmentation (VPS). To study this important problem, we present two datasets, Cityscapes-VPS and VIPER together with a new evaluation metric, video panoptic quality (VPQ). We also propose VPSNet++, an advanced video panoptic segmentation network, which simultaneously performs classification, detection, segmentation, and tracking of all identities in videos. Specifically, VPSNet++ builds upon a top-down panoptic segmentation network by adding pixel-level feature fusion head and object-level association head. The former temporally augments the pixel features while the latter performs object tracking. Furthermore, we propose panoptic boundary learning as an auxiliary task, and instance discrimination learning which learns spatio-temporally clustered pixel embedding for individual thing or stuff regions, i.e., exactly the objective of the video panoptic segmentation problem. Our VPSNet++ significantly outperforms the default VPSNet, i.e., FuseTrack baseline, and achieves state-of-the-art results on both Cityscapes-VPS and VIPER datasets. The datasets, metric, and models are publicly available at https://github.com/mcahny/vps.

Abrupt motion of camera or objects in a scene result in a blurry video, and therefore recovering high quality video requires two types of enhancements: visual enhancement and temporal upsampling. A broad range of research attempted to recover clean frames from blurred image sequences or temporally upsample frames by interpolation, yet there are very limited studies handling both problems jointly. In this work, we present a novel framework for deblurring, interpolating and extrapolating sharp frames from a motion-blurred video in an end-to-end manner. We design our framework by first learning the pixel-level motion that caused the blur from the given inputs via optical flow estimation and then predict multiple clean frames by warping the decoded features with the estimated flows. To ensure temporal coherence across predicted frames and address potential temporal ambiguity, we propose a simple, yet effective flow-based rule. The effectiveness and favorability of our approach are highlighted through extensive qualitative and quantitative evaluations on motion-blurred datasets from high speed videos.

In most of computer vision applications, motion blur is regarded as an undesirable artifact. However, it has been shown that motion blur in an image may have practical interests in fundamental computer vision problems. In this work, we propose a novel framework to estimate optical flow from a single motion-blurred image in an end-to-end manner. We design our network with transformer networks to learn globally and locally varying motions from encoded features of a motion-blurred input, and decode left and right frame features without explicit frame supervision. A flow estimator network is then used to estimate optical flow from the decoded features in a coarse-to-fine manner. We qualitatively and quantitatively evaluate our model through a large set of experiments on synthetic and real motion-blur datasets. We also provide in-depth analysis of our model in connection with related approaches to highlight the effectiveness and favorability of our approach. Furthermore, we showcase the applicability of the flow estimated by our method on deblurring and moving object segmentation tasks.

We present an end-to-end joint training framework that explicitly models 6-DoF motion of multiple dynamic objects, ego-motion, and depth in a monocular camera setup without supervision. Our technical contributions are three-fold. First, we highlight the fundamental difference between inverse and forward projection while modeling the individual motion of each rigid object, and propose a geometrically correct projection pipeline using a neural forward projection module. Second, we design a unified instance-aware photometric and geometric consistency loss that holistically imposes self-supervisory signals for every background and object region. Lastly, we introduce a general-purpose auto-annotation scheme using any off-the-shelf instance segmentation and optical flow models to produce video instance segmentation maps that will be utilized as input to our training pipeline. These proposed elements are validated in a detailed ablation study. Through extensive experiments conducted on the KITTI and Cityscapes dataset, our framework is shown to outperform the state-of-the-art depth and motion estimation methods. Our code, dataset, and models are publicly available.

The booming interest in adversarial attacks stems from a misalignment between human vision and a deep neural network (DNN), \ie~a human imperceptible perturbation fools the DNN. Moreover, a single perturbation, often called universal adversarial perturbation (UAP), can be generated to fool the DNN for most images. A similar misalignment phenomenon has also been observed in the deep steganography task, where a decoder network can retrieve a secret image back from a slightly perturbed cover image. We attempt explaining the success of both in a unified manner from the Fourier perspective. We perform task-specific and joint analysis and reveal that (a) frequency is a key factor that influences their performance based on the proposed entropy metric for quantifying the frequency distribution; (b) their success can be attributed to a DNN being highly sensitive to high-frequency content. We also perform feature layer analysis for providing deep insight on model generalization and robustness. Additionally, we propose two new variants of universal perturbations: (1) high-pass UAP (HP-UAP) being less visible to the human eye; (2) Universal Secret Adversarial Perturbation (USAP) that simultaneously achieves attack and hiding.

The basic framework of depth completion is to predict a pixel-wise dense depth map using very sparse input data. In this paper, we try to solve this problem in a more effective way, by reformulating the regression-based depth estimation problem into a combination of depth plane classification and residual regression. Our proposed approach is to initially densify sparse depth information by figuring out which plane a pixel should lie among a number of discretized depth planes, and then calculate the final depth value by predicting the distance from the specified plane. This will help the network to lessen the burden of directly regressing the absolute depth information from none, and to effectively obtain more accurate depth prediction result with less computation power and inference time. To do so, we firstly introduce a novel way of interpreting depth information with the closest depth plane label p and a residual value r, as we call it, Plane-Residual (PR) representation. We also propose a depth completion network utilizing PR representation consisting of a shared encoder and two decoders, where one classifies the pixel’s depth plane label, while the other one regresses the normalized distance from the classified depth plane. By interpreting depth information in PR representation and using our corresponding depth completion network, we were able to acquire improved depth completion performance with faster computation, compared to previous approaches.

Temporal correspondence - linking pixels or objects across frames - is a fundamental supervisory signal for the video models. For the panoptic understanding of dynamic scenes, we further extend this concept to every segment. Specifically, we aim to learn coarse segment-level matching and fine pixel-level matching together. We implement this idea by designing two novel learning objectives. To validate our proposals, we adopt a deep siamese model and train the model to learn the temporal correspondence on two different levels (i.e., segment and pixel) along with the target task. At inference time, the model processes each frame independently without any extra computation and post-processing. We show that our per-frame inference model can achieve new state-of-the-art results on Cityscapes-VPS and VIPER datasets. Moreover, due to its high efficiency, the model runs in a fraction of time (3×) compared to the previous state-of-the-art approach.

Many advancements in computer vision and machine learning have shown potential for significantly improving the lives of people with disabilities. In particular, recent research has demonstrated that deep neural network models could be used to bridge the gap between the deaf who use sign language and hearing people. The major impediment to advancing such models is the lack of high-quality and large-scale training data. Moreover, previously released sign language datasets include few or no interrogative sentences compared to declarative sentences. In this paper, we introduce a new publicly available large-scale Korean Sign Language (KSL) dataset-KSL-Guide-that includes both declarative sentences and comparable interrogative sentences, which are required for a model to achieve high performance in real-world interactive tasks deployed on service applications. Our dataset contains a total of 121K sign language video samples featuring sentences and words spoken by native KSL speakers with extensive annotations (e.g., gloss, translation, keypoints, and timestamps). We exploit a multi-camera system to produce 3D human pose keypoints as well as 2D keypoints from multi-view RGB. Our experiments quantitatively demonstrate that the inclusion of interrogative sentences in training for sign language recognition and translation tasks greatly improves their performance. Furthermore, we empirically show the qualitative results by developing a prototype application using our dataset, providing an interactive guide service that helps to lower the communication barrier between sign language speakers and hearing people.

Batch normalization (BN) has been widely used in modern deep neural networks (DNNs) due to improved convergence. BN is observed to increase the model accuracy while at the cost of adversarial robustness. There is an increasing interest in the ML community to understand the impact of BN on DNNs, especially related to the model robustness. This work attempts to understand the impact of BN on DNNs from a non-robust feature perspective. Straightforwardly, the improved accuracy can be attributed to the better utilization of useful features. It remains unclear whether BN mainly favors learning robust features (RFs) or non-robust features (NRFs). Our work presents empirical evidence that supports that BN shifts a model towards being more dependent on NRFs. To facilitate the analysis of such a feature robustness shift, we propose a framework for disentangling robust usefulness into robustness and usefulness. Extensive analysis under the proposed framework yields valuable insight on the DNN behavior regarding robustness, e.g. DNNs first mainly learn RFs and then NRFs. The insight that RFs transfer better than NRFs, further inspires simple techniques to strengthen transfer-based black-box attacks.

To reconstruct a 3D scene from a set of calibrated views, traditional multi-view stereo techniques rely on two distinct stages: local depth maps computation and global depth maps fusion. Recent studies concentrate on deep neural architectures for depth estimation by using conventional depth fusion method or direct 3D reconstruction network by regressing Truncated Signed Distance Function (TSDF). In this paper, we advocate that replicating the traditional two stages framework with deep neural networks improves both the interpretability and the accuracy of the results. As mentioned, our network operates in two steps: 1) the local computation of the local depth maps with a deep MVS technique, and, 2) the depth maps and images’ features fusion to build a single TSDF volume. In order to improve the matching performance between images acquired from very different viewpoints (e.g., large-baseline and rotations), we introduce a rotation-invariant 3D convolution kernel called PosedConv. The effectiveness of the proposed architecture is underlined via a large series of experiments conducted on the ScanNet dataset where our approach compares favorably against both traditional and deep learning techniques.

Estimating the motion of the camera together with the 3D structure of the scene from a monocular vision system is a complex task that often relies on the so-called scene rigidity assumption. When observing a dynamic environment, this assumption is violated which leads to an ambiguity between the ego-motion of the camera and the motion of the objects. To solve this problem, we present a self-supervised learning framework for 3D object motion field estimation from monocular videos. Our contributions are two-fold. First, we propose a two-stage projection pipeline to explicitly disentangle the camera ego-motion and the object motions with dynamics attention module, called DAM. Specifically, we design an integrated motion model that estimates the motion of the camera and object in the first and second warping stages, respectively, controlled by the attention module through a shared motion encoder. Second, we propose an object motion field estimation through contrastive sample consensus, called CSAC, taking advantage of weak semantic prior (bounding box from an object detector) and geometric constraints (each object respects the rigid body motion model). Experiments on KITTI, Cityscapes, and Waymo Open Dataset demonstrate the relevance of our approach and show that our method outperforms state-of-the-art algorithms for the tasks of self-supervised monocular depth estimation, object motion segmentation, monocular scene flow estimation, and visual odometry.

Unsupervised Domain Adaptation (UDA) for semantic segmentation has been actively studied to mitigate the domain gap between label-rich source data and unlabeled target data. Despite these efforts, UDA still has a long way to go to reach the fully supervised performance. To this end, we propose a Labeling Only if Required strategy, LabOR, where we introduce a human-in-the-loop approach to adaptively give scarce labels to points that a UDA model is uncertain about. In order to find the uncertain points, we generate an inconsistency mask using the proposed adaptive pixel selector and we label these segment-based regions to achieve near supervised performance with only a small fraction (about 2.2%) ground truth points, which we call "Segment based Pixel-Labeling (SPL)." To further reduce the efforts of the human annotator, we also propose "Point based Pixel-Labeling (PPL)," which finds the most representative points for labeling within the generated inconsistency mask. This reduces efforts from 2.2% segment label → 40 points label while minimizing performance degradation. Through extensive experimentation, we show the advantages of this new framework for domain adaptive semantic segmentation while minimizing human labor costs.

Universal adversarial perturbation (UAP), i.e. a single perturbation to fool the network for most images, is widely recognized as a more practical attack because the UAP can be generated beforehand and applied directly during the at-tack stage. One intriguing phenomenon regarding untargeted UAP is that most images are misclassified to a dominant label. This phenomenon has been reported in previous works while lacking a justified explanation, for which our work attempts to provide an alternative explanation. For a more practical universal attack, our investigation of untargeted UAP focuses on alleviating the dependence on the original training samples, from removing the need for sample labels to limiting the sample size. Towards strictly data-free untargeted UAP, our work proposes to exploit artificial Jigsaw images as the training samples, demonstrating competitive performance. We further investigate the possibility of exploiting the UAP for a data-free black-box attack which is arguably the most practical yet challenging threat model. We demonstrate that there exists optimization-free repetitive patterns which can successfully attack deep models. Code is available at https://bit.ly/3y0ZTIC.

Despite their overwhelming success on a wide range of applications, convolutional neural networks (CNNs) are widely recognized to be vulnerable to adversarial examples. This intriguing phenomenon led to a competition between adversarial attacks and defense techniques. So far, adversarial training is the most widely used method for defending against adversarial attacks. It has also been extended to defend against universal adversarial perturbations (UAPs). The SOTA universal adversarial training (UAT) method optimizes a single perturbation for all training samples in the mini-batch. In this work, we find that a UAP does not attack all classes equally. Inspired by this observation, we identify it as the source of the model having unbalanced robustness. To this end, we improve the SOTA UAT by proposing to utilize class-wise UAPs during adversarial training. On multiple benchmark datasets, our class-wise UAT leads superior performance for both clean accuracy and adversarial robustness against universal attack.

Taking photos and videos with smartphones has become part of our daily life. However, it is still challenging to capture a brief action at the right time (e.g. jump photos), and video editing (e.g. local slow motion) remains a manual, time-consuming task. To address this problem, we present a motion sensor-based approach that leverages the advanced technical features of modern smartphones to facilitate the capture and editing tasks. Concretely, we simultaneously record the motion sensor data from a smartphone carried by the object of interest, as well as a video from a remote smartphone cam-era. Taking advantage of the motion sensor data, our approach can automatically "snap" the video editing effect to the input video, i.e. apply the effect (e.g. local slow motion) at the right time. The proposed approach is shown to be effective in various applications. Moreover, we implemented it as an app for Android smartphones, running in a fully automatic manner.

This paper presents a stereo object matching method that exploits both 2D contextual information from images as well as 3D object-level information. Unlike existing stereo matching methods that exclusively focus on the pixel-level correspondence between stereo images within a volumetric space (i.e., cost volume), we exploit this volumetric structure in a different manner. The cost volume explicitly encompasses 3D information along its disparity axis, therefore it is a privileged structure that can encapsulate the 3D contextual information from objects. However, it is not straightforward since the disparity values map the 3D metric space in a non-linear fashion. Thus, we present two novel strategies to handle 3D objectness in the cost volume space: selective sampling (RoISelect) and 2D-3D fusion (fusion-by-occupancy), which allow us to seamlessly incorporate 3D object-level information and achieve accurate depth performance near the object boundary regions. Our depth estimation achieves competitive performance in the KITTI dataset and the Virtual-KITTI 2.0 dataset.

The intriguing phenomenon of adversarial examples has attracted significant attention in machine learning and what might be more surprising to the community is the existence of universal adversarial perturbations (UAPs), i.e. a single perturbation to fool the target DNN for most images. With the focus on UAP against deep classifiers, this survey summarizes the recent progress on universal adversarial attacks, discussing the challenges from both the attack and defense sides, as well as the reason for the existence of UAP. We aim to extend this work as a dynamic survey that will regularly update its content to follow new works regarding UAP or universal attack in a wide range of domains, such as image, audio, video, text, etc. Relevant updates will be discussed at: https://bit.ly/2SbQlLG. We welcome authors of future works in this field to contact us for including your new findings.

Volumetric deep learning approach towards stereo matching aggregates a cost volume computed from input left and right images using 3D convolutions. Recent works showed that utilization of extracted image features and a spatially varying cost volume aggregation complements 3D convolutions. However, existing methods with spatially varying operations are complex, cost considerable computation time, and cause memory consumption to increase. In this work, we construct Guided Cost volume Excitation (GCE) and show that simple channel excitation of cost volume guided by image can improve performance considerably. Moreover, we propose a novel method of using top-k selection prior to soft-argmin disparity regression for computing the final disparity estimate. Combining our novel contributions, we present an end-to-end network that we call Correlate-and-Excite (CoEx). Extensive experiments of our model on the SceneFlow, KITTI 2012, and KITTI 2015 datasets demonstrate the effectiveness and efficiency of our model and show that our model outperforms other speed-based algorithms while also being competitive to other state-of-the-art algorithms. Codes will be made available at https://github.com/antabangun/coex.

Data hiding is one widely used approach for proving ownership through blind watermarking. Deep learning has been widely used in data hiding, for which inserting an attack simulation layer (ASL) after the watermarked image has been widely recognized as the most effective approach for improving the pipeline robustness against distortions. Despite its wide usage, the gain of enhanced robustness from ASL is usually interpreted through the lens of augmentation, while our work explores this gain from a new perspective by disentangling the forward and backward propagation of such ASL. We find that the main influential component is forward propagation instead of backward propagation. This observation motivates us to use forward ASL to make the pipeline compatible with non-differentiable and/or black-box distortion, such as lossy (JPEG) compression and photoshop effects. Extensive experiments demonstrate the efficacy of our simple approach.

The performance of DNNs trained on clean images has been shown to decrease when the test images have common corruptions. In this work, we interpret corruption robustness as a domain shift and propose to rectify batch normalization (BN) statistics for improving model robustness. This is motivated by perceiving the shift from the clean domain to the corruption domain as a style shift that is represented by the BN statistics. We find that simply estimating and adapting the BN statistics on a few (32 for instance) representation samples, without retraining the model, improves the corruption robustness by a large margin on several benchmark datasets with a wide range of model architectures. For example, on ImageNet-C, statistics adaptation improves the top1 accuracy of ResNet50 from 39.2% to 48.7%. Moreover, we find that this technique can further improve state-of-the-art robust models from 58.1% to 63.3%.

Videos have recently become an omnipresent form of media, gathering much attention from industry as well as academia. In the video enhancement field, video frame interpolation is a long-studied topic that has dramatically improved due to the advancement of deep convolutional neural networks (CNN). However, conventional approaches utilizing two successive frames often exhibit ghosting or tearing artifacts for moving objects. We argue that this phenomenon comes from the lack of reliable information provided only by two frames. With this motivation, we propose a frame interpolation method by utilizing tridirectional information obtained from three input frames. Information extracted from triplet frames allows our model to learn rich and reliable inter-frame motion representations, including subtle nonlinear movement, which can be easily trained via any video frames in a self-supervised manner. We demonstrate that our method generalizes well to high-resolution content by evaluating on FHD resolution, and illustrates our approach's effectiveness via comparison to state-of-the-art methods on challenging video content.

Pursuing a more coherent scene understanding towards real-time vision applications, single-stage instance segmentation has recently gained popularity, achieving a simpler and more efficient design than its two-stage counterparts. Besides, its global mask representation often leads to superior accuracy to the two-stage Mask R-CNN which has been dominant thus far. Despite the promising advances in single-stage methods, finer delineation of instance boundaries still remains unexcavated. Indeed, boundary information provides a strong shape representation that can operate in synergy with the fully-convolutional mask features of the single-stage segmenter. In this work, we propose Boundary Basis based Instance Segmentation(B2Inst) to learn a global boundary representation that can complement existing global-mask-based methods that are often lacking high-frequency details. Besides, we devise a unified quality measure of both mask and boundary and introduce a network block that learns to score the per-instance predictions of itself. When applied to the strongest baselines in single-stage instance segmentation, our B2Inst leads to consistent improvements and accurately parse out the instance boundaries in a scene. Regardless of being single-stage or two-stage frameworks, we outperform the existing state-of-the-art methods on the COCO dataset with the same ResNet-50 and ResNet-101 backbones.

ResNet or DenseNet? Nowadays, most deep learning based approaches are implemented with seminal backbone networks, among them the two arguably most famous ones are ResNet and DenseNet. Despite their competitive performance and overwhelming popularity, inherent drawbacks exist for both of them. For ResNet, the identity shortcut that stabilizes training might limit its representation capacity, and DenseNet mitigates it with multi-layer feature concatenation. However, the dense concatenation causes a new problem of requiring high GPU memory and more training time. Partially due to this, it is not a trivial choice between ResNet and DenseNet. This paper provides a unified perspective of dense summation to analyze them, which facilitates a better understanding of their core difference. We further propose dense weighted normalized shortcuts as a solution to the dilemma between them. Our proposed dense shortcut inherits the design philosophy of simple design in ResNet and DenseNet. On several benchmark datasets, the experimental results show that the proposed DSNet achieves significantly better results than ResNet, and achieves comparable performance as DenseNet but requiring fewer computation resources.

We propose a novel approach to infer a high-quality depth map from a set of images with small viewpoint variations. In general, techniques for depth estimation from small motion consist of camera pose estimation and dense reconstruction. In contrast to prior approaches that recover scene geometry and camera motions using pre-calibrated cameras, we introduce in this paper a self-calibrating bundle adjustment method tailored for small motion which enables computation of camera poses without the need for camera calibration. For dense depth reconstruction, we present a convolutional neural network called DPSNet (Deep Plane Sweep Network) whose design is inspired by best practices of traditional geometry-based approaches. Rather than directly estimating depth or optical flow correspondence from image pairs as done in many previous deep learning methods, DPSNet takes a plane sweep approach that involves building a cost volume from deep features using the plane sweep algorithm, regularizing the cost volume, and regressing the depth map from the cost volume. The cost volume is constructed using a differentiable warping process that allows for end-to-end training of the network. Through the effective incorporation of conventional multiview stereo concepts within a deep learning framework, the proposed method achieves state-of-the-art results on a variety of challenging datasets.

Visual events are usually accompanied by sounds in our daily lives. However, can the machines learn to correlate the visual scene and sound, as well as localize the sound source only by observing them like humans? To investigate its empirical learnability, in this work we first present a novel unsupervised algorithm to address the problem of localizing sound sources in visual scenes. In order to achieve this goal, a two-stream network structure which handles each modality with attention mechanism is developed for sound source localization. The network naturally reveals the localized response in the scene without human annotation. In addition, a new sound source dataset is developed for performance evaluation. Nevertheless, our empirical evaluation shows that the unsupervised method generates false conclusions in some cases. Thereby, we show that this false conclusion cannot be fixed without human prior knowledge due to the well-known correlation and causality mismatch misconception. To fix this issue, we extend our network to the supervised and semi-supervised network settings via a simple modification due to the general architecture of our two-stream network. We show that the false conclusions can be effectively corrected even with a small amount of supervision, i.e., semi-supervised setup. Furthermore, we present the versatility of the learned audio and visual embeddings on the cross-modal content alignment and we extend this proposed algorithm to a new application, sound saliency based automatic camera view panning in 360 degree videos.

For smart manufacturing, an automated robotic assembly system built upon an autoprogramming environment is necessary to reduce setup time and cost for robots that are engaged in frequent task reassignment. This article presents an approach to the autoprogramming of robotic assembly tasks with minimal human assistance. The approach integrates “robotic learning of assembly tasks from observation” and “robotic embodiment of learned assembly tasks in the form of skills.” In the former, robots observe human assembly operations to learn a sequence of assembly tasks, which is formalized into a human assembly script. The latter transforms the human assembly script into a robot assembly script in which a sequence of robot-executable assembly tasks are defined based on action planning supported by workspace modeling and simulated retargeting. The assembly tasks, in the form of the robot assembly script, are then implemented via pretrained robot skills. These skills aim to enable robots to execute difficult tasks that involve inherent uncertainties and variations. We validate the proposed approach by building a prototype of the automated robotic assembly system for a power breaker and an electronic set-top box. The results verify that the proposed automated robotic assembly system is not only feasible but also viable, as it is associated with a dramatic reduction in the human effort required for automating robotic assembly.

A common problem in the task of human-object interaction (HOI) detection is that numerous HOI classes have only a small number of labeled examples, resulting in training sets with a long-tailed distribution. The lack of positive labels can lead to low classification accuracy for these classes. Towards addressing this issue, we observe that there exist natural correlations and anti-correlations among human-object interactions. In this paper, we model the correlations as action co-occurrence matrices and present techniques to learn these priors and leverage them for more effective training, especially on rare classes. The efficacy of our approach is demonstrated experimentally, where the performance of our approach consistently improves over the state-of-the-art methods on both of the two leading HOI detection benchmark datasets, HICO-Det and V-COCO.

Visual storytelling is a task of creating a short story based on photo streams. Unlike existing visual captioning, storytelling aims to contain not only factual descriptions, but also human-like narration and semantics. However, the VIST dataset consists only of a small, fixed number of photos per story. Therefore, the main challenge of visual storytelling is to fill in the visual gap between photos with narrative and imaginative story. In this paper, we propose to explicitly learn to imagine a storyline that bridges the visual gap. During training, one or more photos is randomly omitted from the input stack, and we train the network to produce a full plausible story even with missing photo(s). Furthermore, we propose for visual storytelling a hide-and-tell model, which is designed to learn non-local relations across the photo streams and to refine and improve conventional RNN-based models. In experiments, we show that our scheme of hide-and-tell, and the network design are indeed effective at storytelling, and that our model outperforms previous state-of-the-art methods in automatic metrics. Finally, we qualitatively show the learned ability to interpolate storyline over visual gaps.

A single universal adversarial perturbation (UAP) can be added to all natural images to change most of their predicted class labels. It is of high practical relevance for an attacker to have flexible control over the targeted classes to be attacked, however, the existing UAP method attacks samples from all classes. In this work, we propose a new universal attack method to generate a single perturbation that fools a target network to misclassify only a chosen group of classes, while having limited influence on the remaining classes. Since the proposed attack generates a universal adversarial perturbation that is discriminative to targeted and non-targeted classes, we term it class discriminative universal adversarial perturbation (CD-UAP). We propose one simple yet effective algorithm framework, under which we design and compare various loss function configurations tailored for the class discriminative universal attack. The proposed approach has been evaluated with extensive experiments on various benchmark datasets. Additionally, our proposed approach achieves state-of-the-art performance for the original task of UAP attacking all classes, which demonstrates the effectiveness of our approach.

Panoptic segmentation has become a new standard of visual recognition task by unifying previous semantic segmentation and instance segmentation tasks in concert. In this paper, we propose and explore a new video extension of this task, called video panoptic segmentation. The task requires generating consistent panoptic segmentation as well as an association of instance ids across video frames. To invigorate research on this new task, we present two types of video panoptic datasets. The first is a re-organization of the synthetic VIPER dataset into the video panoptic format to exploit its large-scale pixel annotations. The second is a temporal extension on the Cityscapes val. set, by providing new video panoptic annotations (Cityscapes-VPS). Moreover, we propose a novel video panoptic segmentation network (VPSNet) which jointly predicts object classes, bounding boxes, masks, instance id tracking, and semantic segmentation in video frames. To provide appropriate metrics for this task, we propose a video panoptic quality (VPQ) metric and evaluate our method and several other baselines. Experimental results demonstrate the effectiveness of the presented two datasets. We achieve state-of-the-art results in image PQ on Cityscapes and also in VPQ on Cityscapes-VPS and VIPER datasets.

Convolutional neural network-based approaches have achieved remarkable progress in semantic segmentation. However, these approaches heavily rely on annotated data which are labor intensive. To cope with this limitation, automatically annotated data generated from graphic engines are used to train segmentation models. However, the models trained from synthetic data are difficult to transfer to real images. To tackle this issue, previous works have considered directly adapting models from the source data to the unlabeled target data (to reduce the inter-domain gap). Nonetheless, these techniques do not consider the large distribution gap among the target data itself (intra-domain gap). In this work, we propose a two-step self-supervised domain adaptation approach to minimize the inter-domain and intra-domain gap together. First, we conduct the inter-domain adaptation of the model, from this adaptation, we separate target domain into an easy and hard split using an entropy-based ranking function. Finally, to decrease the intra-domain gap, we propose to employ a self-supervised adaptation technique from the easy to the hard subdomain. Experimental results on numerous benchmark datasets highlight the effectiveness of our method against existing state-of-the-art approaches. The source code is available at https://github.com/feipan664/IntraDA.git.

In this paper, we propose a novel and efficient reference feature extraction module referred to as the Similarity Search and Extraction Network (SSEN) for reference-based super-resolution (RefSR) tasks. The proposed module extracts aligned relevant features from a reference image to increase the performance over single image super-resolution (SISR) methods. In contrast to conventional algorithms which utilize brute-force searches or optical flow estimations, the proposed algorithm is end-to-end trainable without any additional supervision or heavy computation, predicting the best match with a single network forward operation. Moreover, the proposed module is aware of not only the best matching position but also the relevancy of the best match. This makes our algorithm substantially robust when irrelevant reference images are given, overcoming the major cause of the performance degradation when using existing RefSR methods. Furthermore, our module can be utilized for self-similarity SR if no reference image is available. Experimental results demonstrate the superior performance of the proposed algorithm compared to previous works both quantitatively and qualitatively.

A wide variety of works have explored the reason for the existence of adversarial examples, but there is no consensus on the explanation. We propose to treat the DNN logits as a vector for feature representation, and exploit them to analyze the mutual influence of two independent inputs based on the Pearson correlation coefficient (PCC). We utilize this vector representation to understand adversarial examples by disentangling the clean images and adversarial perturbations, and analyze their influence on each other. Our results suggest a new perspective towards the relationship between images and universal perturbations: Universal perturbations contain dominant features, and images behave like noise to them. This feature perspective leads to a new method for generating targeted universal adversarial perturbations using random source images. We are the first to achieve the challenging task of a targeted universal attack without utilizing original training data. Our approach using a proxy dataset achieves comparable performance to the state-of-the-art baselines which utilize the original training dataset.

Taking selfies has become a photographic trend nowadays. We envision the emergence of the "video selfie" capturing a short continuous video clip (or burst photography) of the user, themselves. A selfie stick is usually used, whereby a camera is mounted on a stick for taking selfie photos. In this scenario, we observe that the camera typically goes through a special trajectory along a sphere surface. Motivated by this observation, in this work, we propose an efficient and globally optimal relative camera pose estimation between a pair of two images captured by a camera mounted on a selfie stick. We exploit the special geometric structure of the camera motion constrained by a selfie stick and define its motion as spherical joint motion. By the new parametrization and calibration scheme, we show that the pose estimation problem can be reduced to a 3-DoF (degrees of freedom) search problem, instead of a generic 6-DoF problem. This allows us to derive a fast branch-and-bound global optimization, which guarantees a global optimum. Thereby, we achieve efficient and robust estimation even in the presence of outliers. By experiments on both synthetic and real-world data, we validate the performance as well as the guaranteed optimality of the proposed method.

We present a new linear method for RGB-D based simultaneous localization and mapping (SLAM). Compared to existing techniques relying on the Manhattan world assumption defined by three orthogonal directions, our approach is designed for the more general scenario of the Atlanta world. It consists of a vertical direction and a set of horizontal directions orthogonal to the vertical direction and thus can represent a wider range of scenes. Our approach leverages the structural regularity of the Atlanta world to decouple the non-linearity of camera pose estimations. This allows us separately to estimate the camera rotation and then the translation, which bypasses the inherent non-linearity of traditional SLAM techniques. To this end, we introduce a novel tracking-by-detection scheme to estimate the underlying scene structure by Atlanta representation. Thereby, we propose an Atlanta frame-aware linear SLAM framework which jointly estimates the camera motion and a planar map supporting the Atlanta structure through a linear Kalman filter. Evaluations on both synthetic and real datasets demonstrate that our approach provides favorable performance compared to existing state-of-the-art methods while extending their working range to the Atlanta world.

It is difficult for both cameras and depth sensors to obtain reliable information in hazy scenes. Therefore, image dehazing is still one of the most challenging problems to solve in computer vision and robotics. With the development of convolutional neural networks (CNNs), lots of dehazing and depth estimation algorithms using CNNs have emerged. However, very few of those try to solve these two problems at the same time. Focusing on the fact that traditional haze modeling contains depth information in its formula, we propose a CNN-based simultaneous dehazing and depth estimation network. Our network aims to estimate both a dehazed image and a fully scaled depth map from a single hazy RGB input with end-to-end training. The network contains a single dense encoder and four separate decoders; each of them shares the encoded image representation while performing individual tasks. We suggest a novel depth-transmission consistency loss in the training scheme to fully utilize the correlation between the depth information and transmission map. To demonstrate the robustness and effectiveness of our algorithm, we performed various ablation studies and compared our results to those of state-of-the-art algorithms in dehazing and single image depth estimation, both qualitatively and quantitatively. Furthermore, we show the generality of our network by applying it to some real-world examples.

In this paper, we introduce a new large-scale sidewalk dataset called SideGuide that could potentially help impaired people. Unlike most previous datasets, which are focused on road environments, we paid attention to sidewalks, where understanding the environment could provide the potential for improved walking of humans, especially impaired people. Concretely, we interviewed impaired people and carefully selected target objects from the interviewees' feedback (objects they encounter on sidewalks). We then acquired two different types of data: crowd-sourced data and stereo data. We labeled target objects at instance-level (i.e., bounding box and polygon mask) and generated a ground-truth disparity map for the stereo data. SideGuide consists of 350K images with bounding box annotation, 100K images with a polygon mask, and 180K stereo pairs with the ground-truth disparity. We analyzed our dataset by performing baseline analysis for object detection, instance segmentation, and stereo matching tasks. In addition, we developed a prototype that recognizes the target objects and measures distances, which could potentially assist people with disabilities. The prototype suggests the possibility of practical application of our dataset in real life.

Deep reinforcement learning (DRL) algorithms and evolution strategies (ES) have been applied to various tasks, showing excellent performances. These have the opposite properties, with DRL having good sample efficiency and poor stability, while ES being vice versa. Recently, there have been attempts to combine these algorithms, but these methods fully rely on synchronous update scheme, making it not ideal to maximize the benefits of the parallelism in ES. To solve this challenge, asynchronous update scheme was introduced, which is capable of good time-efficiency and diverse policy exploration. In this paper, we introduce an Asynchronous Evolution Strategy-Reinforcement Learning (AES-RL) that maximizes the parallel efficiency of ES and integrates it with policy gradient methods. Specifically, we propose 1) a novel framework to merge ES and DRL asynchronously and 2) various asynchronous update methods that can take all advantages of asynchronism, ES, and DRL, which are exploration and time efficiency, stability, and sample efficiency, respectively. The proposed framework and update methods are evaluated in continuous control benchmark work, showing superior performance as well as time efficiency compared to the previous methods.

Unsupervised domain adaptation (UDA) for semantic segmentation has been attracting attention recently, as it could be beneficial for various label-scarce real-world scenarios (e.g., robot control, autonomous driving, medical imaging, etc.). Despite the significant progress in this field, current works mainly focus on a single-source single-target setting, which cannot handle more practical settings of multiple targets or even unseen targets. In this paper, we investigate open compound domain adaptation (OCDA), which deals with mixed and novel situations at the same time, for semantic segmentation. We present a novel framework based on three main design principles: discover, hallucinate, and adapt. The scheme first clusters compound target data based on style, discovering multiple latent domains (discover). Then, it hallucinates multiple latent target domains in source by using image-translation (hallucinate). This step ensures the latent domains in the source and the target to be paired. Finally, target-to-source alignment is learned separately between domains (adapt). In high-level, our solution replaces a hard OCDA problem with much easier multiple UDA problems. We evaluate our solution on standard benchmark GTA to C-driving, and achieved new state-of-the-art results.

Neural networks have been shown effective in deep steganography for hiding a full image in another. However, the reason for its success remains not fully clear. Under the existing cover ($C$) dependent deep hiding (DDH) pipeline, it is challenging to analyze how the secret ($S$) image is encoded since the encoded message cannot be analyzed independently. We propose a novel universal deep hiding (UDH) meta-architecture to disentangle the encoding of $S$ from $C$. We perform extensive analysis and demonstrate that the success of deep steganography can be attributed to a frequency discrepancy between $C$ and the encoded secret image. Despite $S$ being hidden in a cover-agnostic manner, strikingly, UDH achieves a performance comparable to the existing DDH. Beyond hiding one image, we push the limits of deep steganography. Exploiting its property of being \emph{universal}, we propose universal watermarking as a timely solution to address the concern of the exponentially increasing amount of images/videos. UDH is robust to a pixel intensity shift on the container image, which makes it suitable for challenging application of light field messaging (LFM). This is the first work demonstrating the success of (DNN-based) hiding a full image for watermarking and LFM. Code: \url{https://github.com/ChaoningZhang/Universal-Deep-Hiding}

We present a simple yet effective prediction module for a one-stage detector. The main process is conducted in a coarse-to-fine manner. First, the module roughly adjusts the default boxes to well capture the extent of target objects in an image. Second, given the adjusted boxes, the module aligns the receptive field of the convolution filters accordingly, not requiring any embedding layers. Both steps build a propose-and-attend mechanism, mimicking two-stage detectors in a highly efficient manner. To verify its effectiveness, we apply the proposed module to a basic one-stage detector SSD. We empirically show that our module significantly lifts the detection accuracy with marginal parameter overhead. Our final model achieves an accuracy comparable to that of state-of-the-art detectors while using a fraction of their model parameter and computational overheads. Moreover, we found that the proposed module has two strong applications. 1) The module can be successfully integrated into a lightweight backbone, further pushing the efficiency of the one-stage detector. 2) The module also allows train-from-scratch without relying on any sophisticated base networks as previous methods do.

Rotating and zooming cameras, also called PTZ (Pan-Tilt-Zoom) cameras, are widely used in modern surveillance systems. While their zooming ability allows acquiring detailed images of the scene, it also makes their calibration more challenging since any zooming action results in a modification of their intrinsic parameters. Therefore, such camera calibration has to be computed online; this process is called self-calibration. In this paper, given an image pair captured by a PTZ camera, we propose a deep learning based approach to automatically estimate the focal length and distortion parameters of both images as well as the rotation angles between them. The proposed approach relies on a dual-Siamese structure, imposing bidirectional constraints. The proposed network is trained on a large-scale dataset automatically generated from a set of panoramas. Empirically, we demonstrate that our proposed approach achieves competitive performance with respect to both deep learning based and traditional state-of-the art methods. Our code and model will be publicly available at https://github.com/ChaoningZhang/DeepPTZ.

While conventional calibrated photometric stereo methods assume that light intensities and sensor exposures are known or unknown but identical across observed images, this assumption easily breaks down in practical settings due to individual light bulb's characteristics and limited control over sensors. This paper studies the effect of unknown and possibly non-uniform light intensities and sensor exposures among observed images on the shape recovery based on photometric stereo. This leads to the development of a "semi-calibrated" photometric stereo method, where the light directions are known but light intensities (and sensor exposures) are unknown. We show that the semi-calibrated photometric stereo becomes a bilinear problem, whose general form is difficult to solve, but in the photometric stereo context, there exists a unique solution for the surface normal and light intensities (or sensor exposures). We further show that there exists a linear solution method for the problem, and develop efficient and stable solution methods. The semi-calibrated photometric stereo is advantageous over conventional calibrated photometric stereo in accurate determination of surface normal, because it relaxes the assumption of known light intensity ratios/sensor exposures. The experimental results show superior accuracy of the semi-calibrated photometric stereo in comparison to conventional methods in practical settings.

In this work, we describe man-made structures via an appropriate structure assumption, called the Atlanta world assumption, which contains a vertical direction (typically the gravity direction) and a set of horizontal directions orthogonal to the vertical direction. Contrary to the commonly used Manhattan world assumption, the horizontal directions in Atlanta world are not necessarily orthogonal to each other. While Atlanta world can encompass a wider range of scenes, this makes the search space much larger and the problem more challenging. Our input data is a set of surface normals, for example, acquired from RGB-D cameras or 3D laser scanners, as well as lines from calibrated images. Given this input data, we propose the first globally optimal method of inlier set maximization for Atlanta direction estimation. We define a novel search space for Atlanta world, as well as its parametrization, and solve this challenging problem using a branch-and-bound (BnB) framework. To alleviate the computational bottleneck in BnB, i.e., the bound computation, we present two bound computation strategies: rectangular bound and slice bound in an efficient measurement domain, i.e., the extended Gaussian image (EGI). In addition, we propose an efficient two-stage method which automatically estimates the number of horizontal directions of a scene. Experimental results with synthetic and real-world datasets have successfully confirmed the validity of our approach.

Video inpainting aims to fill in spatio-temporal holes in videos with plausible content. Despite tremendous progress on deep learning-based inpainting of a single image, it is still challenging to extend these methods to video domain due to the additional time dimension. In this paper, we propose a recurrent temporal aggregation framework for fast deep video inpainting. In particular, we construct an encoder-decoder model, where the encoder takes multiple reference frames which can provide visible pixels revealed from the scene dynamics. These hints are aggregated and fed into the decoder. We apply a recurrent feedback in an auto-regressive manner to enforce temporal consistency in the video results. We propose two architectural designs based on this framework. Our first model is a blind video decaptioning network (BVDNet) that is designed to automatically remove and inpaint text overlays in videos without any mask information. Our BVDNet wins the first place in the ECCV Chalearn 2018 LAP Inpainting Competition Track 2: Video Decaptioning. Second, we propose a network for more general video inpainting (VINet) to deal with more arbitrary and larger holes. Video results demonstrate the advantage of our framework compared to state-of-the-art methods both qualitatively and quantitatively. The codes are available at https://github.com/mcahny/Deep-Video-Inpainting, and https://github.com/shwoo93/video_decaptioning.

Depth from focus (DfF) is a method of estimating the depth of a scene by using information acquired through changes in the focus of a camera. Within the DfF framework of, the focus measure (FM) forms the foundation which determines the accuracy of the output. With the results from the FM, the role of a DfF pipeline is to determine and recalculate unreliable measurements while enhancing those that are reliable. In this paper, we propose a new FM, which we call the "ring difference filter" (RDF), that can more accurately and robustly measure focus. FMs can usually be categorized as confident local methods or noise robust non-local methods. The RDF's unique ring-and-disk structure allows it to have the advantages of both local and non-local FMs. We then describe an efficient pipeline that utilizes the RDF's properties. Part of this pipeline is our proposed RDF-based cost aggregation method, which is able to robustly refine the initial results in the presence of image noise. Our method is able to reproduce results that are on par with or even better than those of state-of-the-art methods, while spending less time in computation.

Video stabilization is a fundamental and important technique for higher quality videos. Prior works have extensively explored video stabilization, but most of them involve cropping of the frame boundaries and introduce moderate levels of distortion. We present a novel deep approach to video stabilization that can generate video frames without cropping and low distortion. The proposed framework utilizes frame interpolation techniques to generate in between frames, leading to reduced inter-frame jitter. Once applied in an iterative fashion, the stabilization effect becomes stronger. A major advantage is that our framework is end-to-end trainable in an unsupervised manner. In addition, our method is able to run in near real-time (15 fps). To the best of our knowledge, this is the first work to propose an unsupervised deep approach to full-frame video stabilization. We show the advantages of our method through quantitative and qualitative evaluations comparing to the state-of-the-art methods.

In this paper, we address the problem of unsupervised video summarization that automatically extracts key-shots from an input video. Specifically, we tackle two critical issues based on our empirical observations: (i) Ineffective feature learning due to flat distributions of output importance scores for each frame, and (ii) training difficulty when dealing with longlength video inputs. To alleviate the first problem, we propose a simple yet effective regularization loss term called variance loss. The proposed variance loss allows a network to predict output scores for each frame with high discrepancy which enables effective feature learning and significantly improves model performance. For the second problem, we design a novel two-stream network named Chunk and Stride Network (CSNet) that utilizes local (chunk) and global (stride) temporal view on the video features. Our CSNet gives better summarization results for long-length videos compared to the existing methods. In addition, we introduce an attention mechanism to handle the dynamic information in videos. We demonstrate the effectiveness of the proposed methods by conducting extensive ablation studies and show that our final model achieves new state-of-the-art results on two benchmark datasets.

Self-supervised tasks such as colorization, inpainting and zigsaw puzzle have been utilized for visual representation learning for still images, when the number of labeled images is limited or absent at all. Recently, this worthwhile stream of study extends to video domain where the cost of human labeling is even more expensive. However, the most of existing methods are still based on 2D CNN architectures that can not directly capture spatio-temporal information for video applications. In this paper, we introduce a new self-supervised task called as Space-Time Cubic Puzzles to train 3D CNNs using large scale video dataset. This task requires a network to arrange permuted 3D spatio-temporal crops. By completing Space-Time Cubic Puzzles, the network learns both spatial appearance and temporal relation of video frames, which is our final goal. In experiments, we demonstrate that our learned 3D representation is well transferred to action recognition tasks, and outperforms state-of-the-art 2D CNN-based competitors on UCF101 and HMDB51 datasets.

Our goal in this work is to train an image captioning model that generates more dense and informative captions. We introduce "relational captioning," a novel image captioning task which aims to generate multiple captions with respect to relational information between objects in an image. Relational captioning is a framework that is advantageous in both diversity and amount of information, leading to image understanding based on relationships. Part-of-speech (POS, i.e. subject-object-predicate categories) tags can be assigned to every English word. We leverage the POS as a prior to guide the correct sequence of words in a caption. To this end, we propose a multi-task triple-stream network (MTTSNet) which consists of three recurrent units for the respective POS and jointly performs POS prediction and captioning. We demonstrate more diverse and richer representations generated by the proposed model against several baselines and competing methods.

In daily life, graphic symbols, such as traffic signs and brand logos, are ubiquitously utilized around us due to its intuitive expression beyond language boundary. We tackle an open-set graphic symbol recognition problem by one-shot classification with prototypical images as a single training example for each novel class. We take an approach to learn a generalizable embedding space for novel tasks. We propose a new approach called variational prototyping-encoder (VPE) that learns the image translation task from real-world input images to their corresponding prototypical images as a meta-task. As a result, VPE learns image similarity as well as prototypical concepts which differs from widely used metric learning based approaches. Our experiments with diverse datasets demonstrate that the proposed VPE performs favorably against competing metric learning based one-shot methods. Also, our qualitative analyses show that our meta-task induces an effective embedding space suitable for unseen data representation.

Blind video decaptioning is a problem of automatically removing text overlays and inpainting the occluded parts in videos without any input masks. While recent deep learning based inpainting methods deal with a single image and mostly assume that the positions of the corrupted pixels are known, we aim at automatic text removal in video sequences without mask information. In this paper, we propose a simple yet effective framework for fast blind video decaptioning. We construct an encoder-decoder model, where the encoder takes multiple source frames that can provide visible pixels revealed from the scene dynamics. These hints are aggregated and fed into the decoder. We apply a residual connection from the input frame to the decoder output to enforce our network to focus on the corrupted regions only. Our proposed model was ranked in the first place in the ECCV Chalearn 2018 LAP Inpainting Competition Track2: Video decaptioning. In addition, we further improve this strong model by applying a recurrent feedback. The recurrent feedback not only enforces temporal coherence but also provides strong clues on where the corrupted pixels are. Both qualitative and quantitative experiments demonstrate that our full model produces accurate and temporally consistent video results in real time (50+ fps).

Video inpainting aims to fill spatio-temporal holes with plausible content in a video. Despite tremendous progress of deep neural networks for image inpainting, it is challenging to extend these methods to the video domain due to the additional time dimension. In this work, we propose a novel deep network architecture for fast video inpainting. Built upon an image-based encoder-decoder model, our framework is designed to collect and refine information from neighbor frames and synthesize still-unknown regions. At the same time, the output is enforced to be temporally consistent by a recurrent feedback and a temporal memory module. Compared with the state-of-the-art image inpainting algorithm, our method produces videos that are much more semantically correct and temporally smooth. In contrast to the prior video completion method which relies on time-consuming optimization, our method runs in near real-time while generating competitive video results. Finally, we applied our framework to video retargeting task, and obtain visually pleasing results.

The performance of deep neural networks improves with more annotated data. The problem is that the budget for annotation is limited. One solution to this is active learning, where a model asks human to annotate data that it perceived as uncertain. A variety of recent methods have been proposed to apply active learning to deep networks but most of them are either designed specific for their target tasks or computationally inefficient for large networks. In this paper, we propose a novel active learning method that is simple but task-agnostic, and works efficiently with the deep networks. We attach a small parametric module, named ``loss prediction module,'' to a target network, and learn it to predict target losses of unlabeled inputs. Then, this module can suggest data that the target model is likely to produce a wrong prediction. This method is task-agnostic as networks are learned from a single loss regardless of target tasks. We rigorously validate our method through image classification, object detection, and human pose estimation, with the recent network architectures. The results demonstrate that our method consistently outperforms the previous methods over the tasks.

Dong-Jin Kim, Jinsoo Choi, Tae-Hyun Oh, In So Kweon. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

In this paper, we present an end-to-end convolutional neural network (CNN) for depth completion. Our network consists of a geometry network and a context network. The geometry network, a single encoder-decoder network, learns to optimize a multi-task loss to generate an initial propagated depth map and a surface normal. The complementary outputs allow it to correctly propagate initial sparse depth points in slanted surfaces. The context network extracts a local and a global feature of an image to compute a bilateral weight, which enables it to preserve edges and fine details in the depth maps. At the end, a final output is produced by multiplying the initially propagated depth map with the bilateral weight. In order to validate the effectiveness and the robustness of our network, we performed extensive ablation studies and compared the results against state-of-the-art CNN-based depth completions, where we showed promising results on various scenes.

In this paper, we present the first large-scale synthetic dataset for visual perception in disaster scenarios, and analyze state-of-the-art methods for multiple computer vision tasks with reference baselines. We simulated before and after disaster scenarios such as fire and building collapse for fifteen different locations in realistic virtual worlds. The dataset consists of more than 300K high-resolution stereo image pairs, all annotated with ground-truth data for semantic segmentation, depth, optical flow, surface normal estimation and camera pose estimation. To create realistic disaster scenes, we manually augmented the effects with 3D models using physical-based graphics tools. We use our dataset to train state-of-the-art methods and evaluate how well these methods can recognize the disaster situations and produce reliable results on virtual scenes as well as real-world images. The results obtained from each task are then used as inputs to the proposed visual odometry network for generating 3D maps of buildings on fire. Finally, we discuss challenges for future research.

As the aging population grows at a rapid rate, there is an ever growing need for service robot platforms that can provide daily assistance at practical speed with reliable performance. In order to assist with daily tasks such as fetching a beverage, a service robot must be able to perceive its environment and generate corresponding motion trajectories. This becomes a challenging and computationally complex problem when the environment is unknown and thus the path planner must sample numerous trajectories that often are sub-optimal, extending the execution time. To address this issue, we propose a unique strategy of integrating a 3D object detection pipeline with a kinematically optimal manipulation planner to significantly increase speed performance at run-time. In addition, we develop a new robotic butler system for a wheeled humanoid that is capable of fetching requested objects at 24% of the speed a human needs to fulfill the same task. The proposed system was evaluated and demonstrated in a real-world environment setup as well as in public exhibition.

We present a method for decomposing the 3D scene flow observed from a moving stereo rig into stationary scene elements and dynamic object motion. Our unsupervised learning framework jointly reasons about the camera motion, optical flow, and 3D motion of moving objects. Three cooperating networks predict stereo matching, camera motion, and residual flow, which represents the flow component due to object motion and not from camera motion. Based on rigid projective geometry, the estimated stereo depth is used to guide the camera motion estimation, and the depth and camera motion are used to guide the residual flow estimation. We also explicitly estimate the 3D scene flow of dynamic objects based on the residual flow and scene depth. Experiments on the KITTI dataset demonstrate the effectiveness of our approach and show that our method outperforms other state-of-the-art algorithms on the optical flow and visual odometry tasks.

In this paper, we present a multi-camera sensor system along with its control algorithm for automated visual inspection from a moving vehicle. To accomplish this task, we propose a unique hardware configuration consisting of a frontal stereo vision system, six lateral cameras motorized to tilt, and a GPS/IMU sensor mounted on the roof of a car. From the frontal stereo system, we detect electric poles and estimate their corresponding 3D positions. Based on this 3D estimation, the tilt angles of the motorized lateral cameras are controlled in real-time to capture high resolution images of the equipment - typically installed a few meters above the road surface. In addition, inertial odometry information from the GPS/IMU module is utilized for pose estimation, object localization, and re-identification among cameras. Experimental results demonstrate the efficiency and robustness of our system for automated electric equipment maintenance, which can reduce human effort significantly.

In this paper, we propose a noise-aware exposure control algorithm for robust robot vision. Our method aims to capture best-exposed images, which can boost the performance of various computer vision and robotics tasks. For this purpose, we carefully design an image quality metric that captures complementary quality attributes and ensures light-weight computation. Specifically, our metric consists of a combination of image gradient, entropy, and noise metrics. The synergy of these measures allows the preservation of sharp edges and rich texture in the image while maintaining a low noise level. Using this novel metric, we propose a real-time and fully automatic exposure and gain control technique based on the Nelder-Mead method. To illustrate the effectiveness of our technique, a large set of experimental results demonstrates the higher qualitative and quantitative performance compared with conventional approaches.

As new display technologies (i.e. foldable phone and modular display) with variable aspect ratios emerge, content-aware video retargeting has attracted much attention from both academia and industry. The content-aware video retargeting aims to adjust the aspect ratio of a video sequence while preserving both, its content and its spatio-temporal consistency. This is a particularly challenging task since these two properties may drastically differ and contradict depending on the video characteristics. In this paper, we explore this conflict in the context of video retargeting, then we propose an appropriate solution to alleviate this issue using a deep recurrent convolutional neural network architecture. First of all, we present a method to generate multiple ground-truth labels under various aspect ratios. Using this dataset, our network is trained to predict various retargeted video candidates from a single input sequence. The resulting candidates present different properties, some of them with more emphasis on the content preservation while the others focus on the spatio-temporal consistency. Among the generated candidates, the final result which satisfy the best compromise is selected. A large set of qualitative and quantitative experiments shows the ability of our method for the content-aware video retargeting.

In this paper, we investigate the problem of unpaired video-to-video translation. Given a video in the source domain, we aim to learn the conditional distribution of the corresponding video in the target domain, without seeing any pairs of corresponding videos. While significant progress has been made in the unpaired translation of images, directly applying these methods to an input video leads to low visual quality due to the additional time dimension. In particular, previous methods suffer from semantic inconsistency (i.e., semantic label flipping) and temporal flickering artifacts. To alleviate these issues, we propose a new framework that is composed of carefully-designed generators and discriminators, coupled with two core objective functions: 1) content preserving loss and 2) temporal consistency loss. Extensive qualitative and quantitative evaluations demonstrate the superior performance of the proposed method against previous approaches. We further apply our framework to a domain adaptation task and achieve favorable results.

Despite recent advances in machine learning, it is still challenging to realize real-time and accurate detection in images. The recently proposed StairNet detector (Sanghyun et al. in Proceedings of winter conference on applications of computer vision (WACV), 2018 ), one of the strongest one-stage detectors, tackles this issue by using a SSD in conjunction with a top-down enrichment module. However, the StairNet approach misses the finer localization information which can be obtained from the lower layer and lacks a feature selection mechanism, which can lead to suboptimal features during the merging step. In this paper, we propose what is termed the gated bidirectional feature pyramid network (GBFPN), a simple and effective architecture that provides a significant improvement over the baseline model, StairNet. The overall network is composed of three parts: a bottom-up pathway , a top-down pathway , and a gating module . Given the multi-scale feature pyramid of deep convolutional network, two separate pathways introduce both finer localization cues and high-level semantics. In each pathway, the gating module dynamically re-weights the features before the combining step, transmitting only the informative features. Placing GBFPN on top of a basic one-stage detector SSD, our method shows state-of-the-art results.

Structure from small motion has become an important topic in 3D computer vision as a method for estimating depth, since capturing the input is so user-friendly. However, major limitations exist with respect to the form of depth uncertainty, due to the narrow baseline and the rolling shutter effect. In this paper, we present a dense 3D reconstruction method from small motion clips using commercial hand-held cameras, which typically cause the undesired rolling shutter artifact. To address these problems, we introduce a novel small motion bundle adjustment that effectively compensates for the rolling shutter effect. Moreover, we propose a pipeline for a fine-scale dense 3D reconstruction that models the rolling shutter effect by utilizing both sparse 3D points and the camera trajectory from narrow-baseline images. In this reconstruction, the sparse 3D points are propagated to obtain an initial depth hypothesis using a geometry guidance term. Then, the depth information on each pixel is obtained by sweeping the plane around each depth search space near the hypothesis. The proposed framework shows accurate dense reconstruction results suitable for various sought-after applications. Both qualitative and quantitative evaluations show that our method consistently generates better depth maps compared to state-of-the-art methods.

One of the core applications of light field imaging is depth estimation. To acquire a depth map, existing approaches apply a single photo-consistency measure to an entire light field. However, this is not an optimal choice because of the non-uniform light field degradations produced by limitations in the hardware design. In this paper, we introduce a pipeline that automatically determines the best configuration for photo-consistency measure, which leads to the most reliable depth label from the light field. We analyzed the practical factors affecting degradation in lenslet light field cameras, and designed a learning based framework that can retrieve the best cost measure and optimal depth label. To enhance the reliability of our method, we augmented an existing light field benchmark to simulate realistic source dependent noise, aberrations, and vignetting artifacts. The augmented dataset was used for the training and validation of the proposed approach. Our method was competitive with several state-of-the-art methods for the benchmark and real-world light field datasets.

Most man-made environments, such as urban and indoor scenes, consist of a set of parallel and orthogonal planar structures. These structures are approximated by the Manhattan world assumption, in which notion can be represented as a Manhattan frame (MF). Given a set of inputs such as surface normals or vanishing points, we pose an MF estimation problem as a consensus set maximization that maximizes the number of inliers over the rotation search space. Conventionally, this problem can be solved by a branch-and-bound framework, which mathematically guarantees global optimality. However, the computational time of the conventional branch-and-bound algorithms is rather far from real-time. In this paper, we propose a novel bound computation method on an efficient measurement domain for MF estimation, i.e., the extended Gaussian image (EGI). By relaxing the original problem, we can compute the bound with a constant complexity, while preserving global optimality. Furthermore, we quantitatively and qualitatively demonstrate the performance of the proposed method for various synthetic and real-world data. We also show the versatility of our approach through three different applications: extension to multiple MF estimation, 3D rotation based video stabilization, and vanishing point estimation (line clustering).

Most artificial neural network models for question-answering rely on complex attention mechanisms. These techniques demonstrate high performance on existing datasets; however, they are limited in their ability to capture natural language variability, and to generate diverse relevant answers. To address this limitation, we propose a model that learns multiple interpretations of a given question. This diversity is ensured by our interpretation policy module which automatically adapts the parameters of a question-answering model with respect to a discrete latent variable. This variable follows the distribution of interpretations learned by the interpretation policy through a semi-supervised variational inference framework. To boost the performance further, the resulting policy is fine-tuned using the rewards from the answer accuracy with a policy gradient. We demonstrate the relevance and efficiency of our model through a large panel of experiments. Qualitative results, in particular, underline the ability of the proposed architecture to discover multiple interpretations of a question. When tested using the Stanford Question Answering Dataset 1.1, our model outperforms the baseline methods in finding multiple and diverse answers. To assess our strategy from a human standpoint, we also conduct a large-scale user study. This study highlights the ability of our network to produce diverse and coherent answers compared to existing approaches. Our Pytorch implementation is available as open source.11github.com/parshakova/APIP.

We introduce a novel method to automatically adjust camera exposure for image processing and computer vision applications on mobile robot platforms. Because most image processing algorithms rely heavily on low-level image features that are based mainly on local gradient information, we consider that gradient quantity can determine the proper exposure level, allowing a camera to capture the important image features in a manner robust to illumination conditions. We then extend this concept to a multi-camera system and present a new control algorithm to achieve both brightness consistency between adjacent cameras and a proper exposure level for each camera. We implement our prototype system with off-the-shelf machine-vision cameras and demonstrate the effectiveness of the proposed algorithms on practical applications, including pedestrian detection, visual odometry, surround-view imaging, panoramic imaging, and stereo matching.

We propose a deep convolutional neural network (CNN) method for natural image matting. Our method takes multiple initial alpha mattes of the previous methods and normalized RGB color images as inputs, and directly learns an end-to-end mapping between the inputs and reconstructed alpha mattes. Among the various existing methods, we focus on using two simple methods as initial alpha mattes: the closed-form matting and KNN matting. They are complementary to each other in terms of local and nonlocal principles. A major benefit of our method is that it can "recognize" different local image structures and then combine the results of local (closed-form matting) and nonlocal (KNN matting) mattings effectively to achieve higher quality alpha mattes than both of the inputs. Furthermore, we verify extendability of the proposed network to different combinations of initial alpha mattes from more advanced techniques such as KL divergence matting and information-flow matting. On the top of deep CNN matting, we build an RGB guided JPEG artifacts removal network to handle JPEG block artifacts in alpha matting. Extensive experiments demonstrate that our proposed deep CNN matting produces visually and quantitatively high-quality alpha mattes. We perform deeper experiments including studies to evaluate the importance of balancing training data and to measure the effects of initial alpha mattes and also consider results from variant versions of the proposed network to analyze our proposed DCNN matting. In addition, our method achieved high ranking in the public alpha matting evaluation dataset in terms of the sum of absolute differences, mean squared errors, and gradient errors. Also, our RGB guided JPEG artifacts removal network restores the damaged alpha mattes from compressed images in JPEG format.

As the computing power of hand-held devices grows, there has been increasing interest in the capture of depth information, to enable a variety of photographic applications. However, under low-light conditions, most devices still suffer from low imaging quality and inaccurate depth acquisition. To address the problem, we present a robust depth estimation method from a short burst shot with varied intensity (i.e., Auto-exposure bracketing) and/or strong noise (i.e., High ISO). Our key idea synergistically combines deep convolutional neural networks with geometric understanding of the scene. We introduce a geometric transformation between optical flow and depth tailored for burst images, enabling our learning-based multi-view stereo matching to be performed effectively. We then describe our depth estimation pipeline that incorporates this geometric transformation into our residual-flow network. It allows our framework to produce an accurate depth map even with a bracketed image sequence. We demonstrate that our method outperforms state-of-the-art methods for various datasets captured by a smartphone and a DSLR camera. Moreover, we show that the estimated depth is applicable for image quality enhancement and photographic editing.

To understand the real-world, it is essential to perceive in all-day conditions including cases which are not suitable for RGB sensors, especially at night. Beyond these limitations, the innovation introduced here is a multispectral solution in the form of depth estimation from a thermal sensor without an additional depth sensor.Based on an analysis of multispectral properties and the relevance to depth predictions, we propose an efficient and novel multi-task framework called the Multispectral Transfer Network (MTN) to estimate a depth image from a single thermal image. By exploiting geometric priors and chromaticity clues, our model can generate a pixel-wise depth image in an unsupervised manner. Moreover, we propose a new type of multitask module called Interleaver as a means of incorporating the chromaticity and fine details of skip-connections into the depth estimation framework without sharing feature layers. Lastly, we explain a novel technical means of stably training and covering large disparities and extending thermal images to data-driven methods for all-day conditions. In experiments, we demonstrate the better performance and generalization of depth estimation through the proposed multispectral stereo dataset, including various driving conditions.

Recent advances in visual recognition show overarching success by virtue of large amounts of supervised data. However, the acquisition of a large supervised dataset is often challenging. This is also true for intelligent transportation applications, i.e., traffic sign recognition. For example, a model trained with data of one country may not be easily generalized to another country without much data. We propose a novel feature embedding scheme for unseen class classification when the representative class template is given. Traffic signs, unlike other objects, have official images. We perform co-domain embedding using a quadruple relationship from real and synthetic domains. Our quadruplet network fully utilizes the explicit pairwise similarity relationships among samples from different domains. We validate our method on three datasets with two experiments involving one-shot classification and feature generalization. The results show that the proposed method outperforms competing approaches on both seen and unseen classes.

As demand for advanced photographic applications on hand-held devices grows, these electronics require the capture of high quality depth. However, under low-light conditions, most devices still suffer from low imaging quality and inaccurate depth acquisition. To address the problem, we present a robust depth estimation method from a short burst shot with varied intensity (i.e., Auto Bracketing) or strong noise (i.e., High ISO). We introduce a geometric transformation between flow and depth tailored for burst images, enabling our learning-based multi-view stereo matching to be performed effectively. We then describe our depth estimation pipeline that incorporates the geometric transformation into our residual-flow network. It allows our framework to produce an accurate depth map even with a bracketed image sequence. We demonstrate that our method outperforms state-of-the-art methods for various datasets captured by a smartphone and a DSLR camera. Moreover, we show that the estimated depth is applicable for image quality enhancement and photographic editing.

In this work, we describe man-made structures via an appropriate structure assumption, called Atlanta world, which contains a vertical direction (typically the gravity direction) and a set of horizontal directions orthogonal to the vertical direction. Contrary to the commonly used Manhattan world assumption, the horizontal directions in Atlanta world are not necessarily orthogonal to each other. While Atlanta world permits to encompass a wider range of scenes, this makes the solution space larger and the problem more challenging. Given a set of inputs, such as lines in a calibrated image or surface normals, we propose the first globally optimal method of inlier set maximization for Atlanta direction estimation. We define a novel search space for Atlanta world, as well as its parameterization, and solve this challenging problem by a branch-and-bound framework. Experimental results with synthetic and real-world datasets have successfully confirmed the validity of our approach.

Learning-based color enhancement approaches typically learn to map from input images to retouched images. Most of existing methods require expensive pairs of input-retouched images or produce results in a non-interpretable way. In this paper, we present a deep reinforcement learning (DRL) based method for color enhancement to explicitly model the step-wise nature of human retouching process. We cast a color enhancement process as a Markov Decision Process where actions are defined as global color adjustment operations. Then we train our agent to learn the optimal global enhancement sequence of the actions. In addition, we present a 'distort-and-recover' training scheme which only requires high-quality reference images for training instead of input and retouched image pairs. Given high-quality reference images, we distort the images' color distribution and form distorted-reference image pairs for training. Through extensive experiments, we show that our method produces decent enhancement results and our DRL approach is more suitable for the 'distort-and-recover' training scheme than previous supervised approaches. Supplementary material and code are available at https://sites.google.com/view/distort-and-recover/.

Visual events are usually accompanied by sounds in our daily lives. We pose the question: Can the machine learn the correspondence between visual scene and the sound, and localize the sound source only by observing sound and visual scene pairs like human? In this paper, we propose a novel unsupervised algorithm to address the problem of localizing the sound source in visual scenes. A two-stream network structure which handles each modality, with attention mechanism is developed for sound source localization. Moreover, although our network is formulated within the unsupervised learning framework, it can be extended to a unified architecture with a simple modification for the supervised and semi-supervised learning settings as well. Meanwhile, a new sound source dataset is developed for performance evaluation. Our empirical evaluation shows that the unsupervised method eventually go through false conclusion in some cases. We also show that even with a few supervision, i.e., semi-supervised setup, false conclusion is able to be corrected effectively.

Light field cameras capture both the spatial and the angular properties of light rays in space. Due to its property, one can compute the depth from light fields in uncontrolled lighting environments, which is a big advantage over active sensing devices. Depth computed from light fields can be used for many applications including 3D modelling and refocusing. However, light field images from hand-held cameras have very narrow baselines with noise, making the depth estimation difficult. Many approaches have been proposed to overcome these limitations for the light field depth estimation, but there is a clear trade-off between the accuracy and the speed in these methods. In this paper, we introduce a fast and accurate light field depth estimation method based on a fully-convolutional neural network. Our network is designed by considering the light field geometry and we also overcome the lack of training data by proposing light field specific data augmentation methods. We achieved the top rank in the HCI 4D Light Field Benchmark on most metrics, and we also demonstrate the effectiveness of the proposed method on real-world light-field images.

Objects and their relationships are critical contents for image understanding. A scene graph provides a structured description that captures these properties of an image. However, reasoning about the relationships between objects is very challenging and only a few recent works have attempted to solve the problem of generating a scene graph from an image. In this paper, we present a novel method that improves scene graph generation by explicitly modeling inter-dependency among the entire object instances. We design a simple and effective relational embedding module that enables our model to jointly represent connections among all related objects, rather than focus on an object in isolation. Our novel method significantly benefits two main parts of the scene graph generation task: object classification and relationship classification. Using it on top of a basic Faster R-CNN, our model achieves state-of-the-art results on the Visual Genome benchmark. We further push the performance by introducing global context encoding module and geometrical layout encoding module. We validate our final model, LinkNet, through extensive ablation studies, demonstrating its efficacy in scene graph generation.

Human behavior understanding is arguably one of the most important mid-level components in artificial intelligence. In order to efficiently make use of data, multi-task learning has been studied in diverse computer vision tasks including human behavior understanding. However, multitask learning relies on task specific datasets and constructing such datasets can be cumbersome. It requires huge amounts of data, labeling efforts, statistical consideration etc. In this paper, we leverage existing single-task datasets for human action classification and captioning data for efficient human behavior learning. Since the data in each dataset has respective heterogeneous annotations, traditional multi-task learning is not effective in this scenario. To this end, we propose a novel alternating directional optimization method to efficiently learn from the heterogeneous data. We demonstrate the effectiveness of our model and show performance improvements on both classification and sentence retrieval tasks in comparison to the models trained on each of the single-task datasets.

The best summary of a long video differs among different people due to its highly subjective nature. Even for the same person, the best summary may change with time or mood. In this paper, we introduce the task of generating contextually customized video summaries through simple text. First, we train a deep architecture to effectively learn semantic embeddings of video frames by leveraging the abundance of image-caption data via a progressive manner, whereby our algorithm is able to select semantically relevant video segments for a contextually meaningful video summary, given a user-specific text description or even a single sentence. In order to evaluate our customized video summaries, we conduct experimental comparison with baseline methods that utilize ground-truth information. Despite the challenging baselines, our method still manages to show comparable or even exceeding performance. We also demonstrate that our method is able to automatically generate semantically diverse video summaries even without any text input.

In this paper, we explore methods of complicating selfsupervised tasks for representation learning. That is, we do severe damage to data and encourage a network to recover them. First, we complicate each of three powerful self-supervised task candidates: jigsaw puzzle, inpainting, and colorization. In addition, we introduce a novel complicated self-supervised task called "Completing damaged jigsaw puzzles" which is puzzles with one piece missing and the other pieces without color. We train a convolutional neural network not only to solve the puzzles, but also generate the missing content and colorize the puzzles. The recovery of the aforementioned damage pushes the network to obtain robust and general-purpose representations. We demonstrate that complicating the self-supervised tasks improves their original versions and that our final task learns more robust and transferable representations compared to the previous methods, as well as the simple combination of our candidate tasks. Our approach achieves state-of-the-art performance in transfer learning on PASCAL classification and semantic segmentation.

One-stage object detectors such as SSD or YOLO already have shown promising accuracy with small memory footprint and fast speed. However, it is widely recognized that one-stage detectors have difficulty in detecting small objects while they are competitive with two-stage methods on large objects. In this paper, we investigate how to alleviate this problem starting from the SSD framework. Due to their pyramidal design, the lower layer that is responsible for small objects lacks strong semantics(e.g contextual information). We address this problem by introducing a feature combining module that spreads out the strong semantics in a top-down manner. Our final model StairNet detector unifies the multi-scale representations and semantic distribution effectively. Experiments on PASCAL VOC 2007 and PASCAL VOC 2012 datasets demonstrate that Stair-Net significantly improves the weakness of SSD and outperforms the other state-of-the-art one-stage detectors.

Rank minimization can be converted into tractable surrogate problems, such as Nuclear Norm Minimization (NNM) and Weighted NNM (WNNM). The problems related to NNM, or WNNM, can be solved iteratively by applying a closed-form proximal operator, called Singular Value Thresholding (SVT), or Weighted SVT, but they suffer from high computational cost of Singular Value Decomposition (SVD) at each iteration. We propose a fast and accurate approximation method for SVT, that we call fast randomized SVT (FRSVT), with which we avoid direct computation of SVD. The key idea is to extract an approximate basis for the range of the matrix from its compressed matrix. Given the basis, we compute partial singular values of the original matrix from the small factored matrix. In addition, by developping a range propagation method, our method further speeds up the extraction of approximate basis at each iteration. Our theoretical analysis shows the relationship between the approximation bound of SVD and its effect to NNM via SVT. Along with the analysis, our empirical results quantitatively and qualitatively show that our approximation rarely harms the convergence of the host algorithms. We assess the efficiency and accuracy of the proposed method on various computer vision problems, e.g., subspace clustering, weather artifact removal, and simultaneous multi-image alignment and rectification.

The introduction of small motion techniques such as small angle rotation approximation has enabled the three-dimensional reconstruction from a small motion of a camera, so-called structure from small motion (SfSM). In this letter, we propose a closed-form solution dedicated to the rotation estimation problem in SfSM. We show that our method works with a minimal set of two points, and has mild conditions to produce a unique optimal solution in practice. Also, we introduce a three-step SfSM pipeline with better convergence and faster speed compared to the state-of-the-art SfSM approaches. The key to this improvement is the separated estimation of the rotation with the proposed two-point method in order to handle the bas-relief ambiguity that affects the convergence of the bundle adjustment. We demonstrate the effectiveness of our two-point minimal solution and the three-step SfSM approach in synthetic and real-world experiments under the small motion regime.

Inertial measurement units (IMUs) are successfully utilized to compensate localization errors in sensor fused inertial navigation systems. An IMU generally produces high-frequency signals ranging from 100 to 1000 Hz, and preintegration methods are applied to effectively process these high-frequency signals for inertial navigation systems. The main problem with an existing preintegration method is that the inertial propagation models in the method are only generated at the IMU's coordinate system. Hence, the models have to be converted to the coordinate system of the other sensor in order to apply its constraint. So, the iterative optimization framework using the conventional method takes large amount of time. In addition, since a general rigid body transformation cannot transfer a velocity propagation model to the other coordinate system, the concept of relative motion analysis needs to be considered. To solve the problems above, in this paper, we propose a novel relative preintegration (RP) method that can generate inertial propagation models at any sensor's coordinate system in a rigid body. This permits accurate and fast IMU processing in sensor fused inertial navigation systems. We applied new nonlinear optimization frameworks to solve initialization and extrinsic calibration problems for the IMU-IMU, IMU-Camera, and IMU-LiDAR pair based on the proposed RP method in an on-line manner, and the superior results of the mentioned processes are presented as well.

We introduce the KAIST multi-spectral data set, which covers a great range of drivable regions, from urban to residential, for autonomous systems. Our data set provides the different perspectives of the world captured in coarse time slots (day and night), in addition to fine time slots (sunrise, morning, afternoon, sunset, night, and dawn). For all-day perception of autonomous systems, we propose the use of a different spectral sensor, i.e., a thermal imaging camera. Toward this goal, we develop a multi-sensor platform, which supports the use of a co-aligned RGB/Thermal camera, RGB stereo, 3-D LiDAR, and inertial sensors (GPS/IMU) and a related calibration technique. We design a wide range of visual perception tasks including the object detection, drivable region detection, localization, image enhancement, depth estimation, and colorization using a single/multi-spectral approach. In this paper, we provide a description of our benchmark with the recording platform, data format, development toolkits, and lessons about the progress of capturing data sets.

In this paper, we introduce robust and synergetic hand-crafted features and a simple but efficient deep feature from a convolutional neural network (CNN) architecture for defocus estimation. This paper systematically analyzes the effectiveness of different features, and shows how each feature can compensate for the weaknesses of other features when they are concatenated. For a full defocus map estimation, we extract image patches on strong edges sparsely, after which we use them for deep and hand-crafted feature extraction. In order to reduce the degree of patch-scale dependency, we also propose a multi-scale patch extraction strategy. A sparse defocus map is generated using a neural network classifier followed by a probability-joint bilateral filter. The final defocus map is obtained from the sparse defocus map with guidance from an edge-preserving filtered input image. Experimental results show that our algorithm is superior to state-of-the-art algorithms in terms of defocus estimation. Our work can be used for applications such as segmentation, blur magnification, all-in-focus image generation, and 3-D estimation.

Depth from focus (DfF) is a method of estimating depth of a scene by using the information acquired through the change of the focus of a camera. Within the framework of DfF, the focus measure (FM) forms the foundation on which the accuracy of the output is determined. With the result from the FM, the role of a DfF pipeline is to determine and recalculate unreliable measurements while enhancing those that are reliable. In this paper, we propose a new FM that more accurately and robustly measures focus, which we call the ring difference filter (RDF). FMs can usually be categorized as confident local methods or noise robust non-local methods. RDFs unique ring-and-disk structure allows it to have the advantageous sides of both local and non-local FMs. We then describe an efficient pipeline that utilizes the properties that the RDF brings. Our method is able to reproduce results that are on par with or even better than those of the state-of-the-art, while spending less time in computation.

This paper proposes a weakly- and self-supervised deep convolutional neural network (WSSDCNN) for content-aware image retargeting. Our network takes a source image and a target aspect ratio, and then directly outputs a retargeted image. Retargeting is performed through a shift reap, which is a pixel-wise mapping from the source to the target grid. Our method implicitly learns an attention map, which leads to r content-aware shift map for image retargeting. As a result, discriminative parts in an image are preserved, while background regions are adjusted seamlessly. In the training phase, pairs of an image and its image-level annotation are used to compute content and structure tosses. We demonstrate the effectiveness of our proposed method for a retargeting application with insightful analyses.

The recent proliferation of high resolution cameras presents an opportunity to achieve unprecedented levels of precision in visual 3D reconstruction. Yet the camera calibration pipeline, developed decades ago using checkerboards, has remained the de facto standard. In this paper, we ask the question: are checkerboards the optimal pattern for high precision calibration? We empirically demonstrate that deltille grids (regular triangular tiling) produce the highest precision calibration of the possible tilings of Euclidean plane. We posit that they should be the new standard for high-precision calibration and present a complete ecosystem for calibration using deltille grids including: (1) a highly precise corner detection algorithm based on polynomial surface fitting; (2) an indexing scheme based on polarities extracted from the fitted surfaces; and (3) a 2D coding system for deltille grids, which we refer to as DelTags, in lieu of conventional matrix barcodes. We demonstrate state-of-the-art performance and apply the full calibration ecosystem through the use of 3D calibration objects for multiview camera calibration.

In this paper, we propose a unified end-to-end trainable multi-task network that jointly handles lane and road marking detection and recognition that is guided by a vanishing point under adverse weather conditions. We tackle rainy and low illumination conditions, which have not been extensively studied until now due to clear challenges. For example, images taken under rainy days are subject to low illumination, while wet roads cause light reflection and distort the appearance of lane and road markings. At night, color distortion occurs under limited illumination. As a result, no benchmark dataset exists and only a few developed algorithms work under poor weather conditions. To address this shortcoming, we build up a lane and road marking benchmark which consists of about 20,000 images with 17 lane and road marking classes under four different scenarios: no rain, rain, heavy rain, and night. We train and evaluate several versions of the proposed multi-task network and validate the importance of each task. The resulting approach, VPGNet, can detect and classify lanes and road markings, and predict a vanishing point with a single forward pass. Experimental results show that our approach achieves high accuracy and robustness under various conditions in realtime (20 fps). The benchmark and the VPGNet model will be publicly available.

Cinemagraphs are a compelling way to convey dynamic aspects of a scene. In these media, dynamic and still elements are juxtaposed to create an artistic and narrative experience. Creating a high-quality, aesthetically pleasing cinemagraph requires isolating objects in a semantically meaningful way and then selecting good start times and looping periods for those objects to minimize visual artifacts (such a tearing). To achieve this, we present a new technique that uses object recognition and semantic segmentation as part of an optimization method to automatically create cinemagraphs from videos that are both visually appealing and semantically meaningful. Given a scene with multiple objects, there are many cinemagraphs one could create. Our method evaluates these multiple candidates and presents the best one, as determined by a model trained to predict human preferences in a collaborative way. We demonstrate the effectiveness of our approach with multiple results and a user study.

We propose a novel video object segmentation algorithm based on pixel-level matching using Convolutional Neural Networks (CNN). Our network aims to distinguish the target area from the background on the basis of the pixel-level similarity between two object units. The proposed network represents a target object using features from different depth layers in order to take advantage of both the spatial details and the category-level semantic information. Furthermore, we propose a feature compression technique that drastically reduces the memory requirements while maintaining the capability of feature representation. Two-stage training (pretraining and fine-tuning) allows our network to handle any target object regardless of its category (even if the object's type does not belong to the pre-training data) or of variations in its appearance through a video sequence. Experiments on large datasets demonstrate the effectiveness of our model - against related methods - in terms of accuracy, speed, and stability. Finally, we introduce the transferability of our network to different domains, such as the infrared data domain.

In this paper, we present a visual learning framework to retrieve a 3D model and estimate its pose from a single image. To increase the quantity and quality of training data, we define our simulation space in the near infrared (NIR) band, and utilize the quasi-Monte Carlo (MC) method for scalable photorealistic rendering of manufactured components. Two types of convolutional neural network (CNN) architectures are trained over these synthetic data and a relatively small amount of real data. The first CNN model seeks the most discriminative information and uses it to classify industrial components with fine-grained shape attributes. Once a 3D model is identified, one of the category-specific CNNs is tested for pose regression in the second phase. The mixed data for learning object categories is useful in domain adaptation and attention mechanism in our system. We validate our data-driven method with 88 component models, and the experimental results are qualitatively demonstrated. Also, the CNNs trained with various conditions of mixed data are quantitatively analyzed.

This paper presents a robust approach for road marking detection and recognition from images captured by an embedded camera mounted on a car. Our method is designed to cope with illumination changes, shadows, and harsh meteorological conditions. Furthermore, the algorithm can effectively group complex multi-symbol shapes into an individual road marking. For this purpose, the proposed technique relies on MSER features to obtain candidate regions which are further merged using density-based clustering. Finally, these regions of interest are recognized using machine learning approaches. Worth noting, the algorithm is versatile since it does not utilize any prior information about lane position or road space. The proposed method compares favorably to other existing works through a large number of experiments on an extensive road marking dataset.

Abstract In this paper, we present a new framework to determine up front orientations and detect salient views of 3D models. The salient viewpoint to human preferences is the most informative projection with correct upright orientation. Our method utilizes two Convolutional Neural Network (CNN) architectures to encode category‐specific information learnt from a large number of 3D shapes and 2D images on the web. Using the first CNN model with 3D voxel data, we generate a CNN shape feature to decide natural upright orientation of 3D objects. Once a 3D model is upright‐aligned, the front projection and salient views are scored by category recognition using the second CNN model. The second CNN is trained over popular photo collections from internet users. In order to model comfortable viewing angles of 3D models, a category‐dependent prior is also learnt from the users. Our approach effectively combines category‐specific scores and classical evaluations to produce a data‐driven viewpoint saliency map. The best viewpoints from the method are quantitatively and qualitatively validated with more than 100 objects from 20 categories. Our thumbnail images of 3D models are the most favoured among those from different approaches.

We present a novel method for the geometric calibration of micro-lens-based light field cameras. Accurate geometric calibration is the basis of various applications. Instead of using sub-aperture images, we directly utilize raw images for calibration. We select appropriate regions in raw images and extract line features from micro-lens images in those regions. For the entire process, we formulate a new projection model of a micro-lens-based light field camera, which contains a smaller number of parameters than previous models. The model is transformed into a linear form using line features. We compute the initial solution of both the intrinsic and the extrinsic parameters by a linear computation and refine them via non-linear optimization. Experimental results demonstrate the accuracy of the correspondences between rays and pixels in raw images, as estimated by the proposed method.

In this paper, we introduce an automatic approach to generate trimaps and consistent alpha mattes of foreground objects in a light-field image. Our method first performs binary segmentation to roughly segment a light-field image into foreground and background based on depth and color. Next, we estimate accurate trimaps through analyzing color distribution along the boundary of the segmentation using guided image filter and KL-divergence. In order to estimate consistent alpha mattes across sub-images, we utilize the epipolar plane image (EPI) where colors and alphas along the same epipolar line must be consistent. Since EPI of foreground and background are mixed in the matting area, we propagate the EPI from definite foreground/background regions to unknown regions by assuming depth variations within unknown regions are spatially smooth. Using the EPI constraint, we derive two solutions to estimate alpha when color samples along epipolar line are known, and unknown. To further enhance consistency, we refine the estimated alpha mattes by using the multi-image matting Laplacian with an additional EPI smoothness constraint. In experimental evaluations, we have created a dataset where the ground truth alpha mattes of light-field images were obtained by using the blue screen technique. A variety of experiments show that our proposed algorithm produces both visually and quantitatively high-quality alpha mattes for light-field images.

We propose a robust uncalibrated multiview photometric stereo method for high quality 3D shape reconstruction. In our method, a coarse initial 3D mesh obtained using a multiview stereo method is projected onto a 2D planar domain using a planar mesh parameterization technique. We describe methods for surface normal estimation that work in the parameterized 2D space that jointly incorporates all geometric and photometric cues from multiple viewpoints. Using an estimated surface normal map, a refined 3D mesh is then recovered by computing an optimal displacement map in the same 2D planar domain. Our method avoids the need of merging view-dependent surface normal maps that is often required in conventional methods. We conduct evaluation on various real-world objects containing surfaces with specular reflections, multiple albedos, and complex topologies in both controlled and uncontrolled settings and demonstrate that accurate 3D meshes with fine geometric details can be recovered by our method.

In this letter, we present an accurate Depth from Small Motion approach, which reconstructs three-dimensional (3-D) depth from image sequences with extremely narrow baselines. We start with estimating sparse 3-D points and camera poses via the structure from motion method. For dense depth reconstruction, we propose a novel depth propagation using a geometric guidance term that considers not only the geometric constraint from the surface normal, but also color consistency. In addition, we propose an accurate surface normal estimation method with a multiple range search so that the normal vector can guide the direction of the depth propagation precisely. The major benefit of our depth propagation method is that it obtains detailed structures of a scene without fronto-parallel bias. We validate our method using various indoor and outdoor datasets, and both qualitative and quantitative experimental results show that our new algorithm consistently generates better 3-D depth information than the results of existing state-of-the-art methods.

Commercial light field cameras provide spatial and angular information, but their limited resolution becomes an important problem in practical use. In this letter, we present a novel method for light field image super-resolution (SR) to simultaneously up-sample both the spatial and angular resolutions of a light field image via a deep convolutional neural network. We first augment the spatial resolution of each subaperture image by a spatial SR network, then novel views between super-resolved subaperture images are generated by three different angular SR networks according to the novel view locations. We improve both the efficiency of training and the quality of angular SR results by using weight sharing. In addition, we provide a new light field image dataset for training and validating the network. We train our whole network end-to-end, and show state-of-the-art performances on quantitative and qualitative evaluations.

Nearly all 3D displays need calibration for correct rendering. More often than not, the optical elements in a 3D display are misaligned from the designed parameter setting. As a result, 3D magic does not perform well as intended. The observed images tend to get distorted. In this paper, we propose a novel display calibration method to fix the situation. In our method, a pattern image is displayed on the panel and a camera takes its pictures twice at different positions. Then, based on a quantitative model, we extract all display parameters (i.e., pitch, slanted angle, gap or thickness, and offset) from the observed patterns in the captured images. For high accuracy and robustness, our method analyzes the patterns mostly in the frequency domain. We conduct two types of experiments for validation; one with optical simulation for quantitative results and the other with real-life displays for qualitative assessment. Experimental results demonstrate that our method is quite accurate, about a half order of magnitude higher than prior work; is efficient, spending less than 2s for computation; and is robust to noise, working well in the SNR regime as low as 6dB.

We present a novel coded exposure video technique for multi-image motion deblurring. The key idea of this paper is to capture video frames with a set of complementary fluttering patterns, which enables us to preserve all spectrum bands of a latent image and recover a sharp latent image. To achieve this, we introduce an algorithm for generating a complementary set of binary sequences based on the modern communication theory and implement the coded exposure video system with an off-the-shelf machine vision camera. To demonstrate the effectiveness of our method, we provide in-depth analyses of the theoretical bounds and the spectral gains of our method and other state-of-the-art computational imaging approaches. We further show deblurring results on various challenging examples with quantitative and qualitative comparisons to other computational image capturing methods used for image deblurring, and show how our method can be applied for protecting privacy in videos.

Near-Infrared (NIR) images of most materials exhibit less texture or albedo variations making them beneficial for vision tasks such as intrinsic image decomposition and structured light depth estimation. Understanding the reflectance properties (BRDF) of materials in the NIR wavelength range can be further useful for many photometric methods including shape from shading and inverse rendering. However, even with less albedo variation, many materials e.g. fabrics, leaves, etc. exhibit complex fine-scale surface detail making it hard to accurately estimate BRDF. In this paper, we present an approach to simultaneously estimate NIR BRDF and fine-scale surface details by imaging materials under different IR lighting and viewing directions. This is achieved by an iterative scheme that alternately estimates surface detail and NIR BRDF of materials. Our setup does not require complicated gantries or calibration and we present the first NIR dataset of 100 materials including a variety of fabrics (knits, weaves, cotton, satin, leather), and organic (skin, leaves, jute, trunk, fur) and inorganic materials (plastic, concrete, carpet). The NIR BRDFs measured from material samples are used with a shape-from-shading algorithm to demonstrate fine-scale reconstruction of objects from a single NIR image.

We address the problem of composing a story out of multiple short video clips taken by a person during an activity or experience. Inspired by plot analysis of written stories, our method generates a sequence of video clips ordered in such a way that it reflects plot dynamics and content coherency. That is, given a set of multiple video clips, our method composes a video which we call a video-story. We define metrics on scene dynamics and coherency by dense optical flow features and a patch matching algorithm. Using these metrics, we define an objective function for the video-story. To efficiently search for the best video-story, we introduce a novel Branch-and-Bound algorithm which guarantees the global optimum. We collect the dataset consisting of 23 video sets from the web, resulting in a total of 236 individual video clips. With the acquired dataset, we perform extensive user studies involving 30 human subjects by which the effectiveness of our approach is quantitatively and qualitatively verified.

We propose a novel approach that generates a highquality depth map from a set of images captured with a small viewpoint variation, namely small motion clip. As opposed to prior methods that recover scene geometry and camera motions using pre-calibrated cameras, we introduce a self-calibrating bundle adjustment tailored for small motion. This allows our dense stereo algorithm to produce a high-quality depth map for the user without the need for camera calibration. In the dense matching, the distributions of intensity profiles are analyzed to leverage the benefit of having negligible intensity changes within the scene due to the minuscule variation in viewpoint. The depth maps obtained by the proposed framework show accurate and extremely fine structures that are unmatched by previous literature under the same small motion configuration.

Consumer devices with stereo cameras have become popular because of their low-cost depth sensing capability. However, those systems usually suffer from low imaging quality and inaccurate depth acquisition under low-light conditions. To address the problem, we present a new stereo matching method with a color and monochrome camera pair. We focus on the fundamental trade-off that monochrome cameras have much better light-efficiency than color-filtered cameras. Our key ideas involve compensating for the radiometric difference between two cross-spectral images and taking full advantage of complementary data. Consequently, our method produces both an accurate depth map and high-quality images, which are applicable for various depth-aware image processing. Our method is evaluated using various datasets and the performance of our depth estimation consistently outperforms state-of-the-art methods.

Given a set of surface normals, we pose a Manhattan Frame (MF) estimation problem as a consensus set maximization that maximizes the number of inliers over the rotation search space. We solve this problem through a branchand-bound framework, which mathematically guarantees a globally optimal solution. However, the computational time of conventional branch-and-bound algorithms are intractable for real-time performance. In this paper, we propose a novel bound computation method within an efficient measurement domain for MF estimation, i.e., the extended Gaussian image (EGI). By relaxing the original problem, we can compute the bounds in real-time, while preserving global optimality. Furthermore, we quantitatively and qualitatively demonstrate the performance of the proposed method for synthetic and real-world data. We also show the versatility of our approach through two applications: extension to multiple MF estimation and video stabilization.

We introduce a new technique that automatically generates diverse, visually compelling stylizations for a photograph in an unsupervised manner. We achieve this by learning style ranking for a given input using a large photo collection and selecting a diverse subset of matching styles for final style transfer. We also propose an improved technique that transfers the global color and tone of the chosen exemplars to the input photograph while avoiding the common visual artifacts produced by the existing style transfer methods. Together, our style selection and transfer techniques produce compelling, artifact-free results on a wide range of input photographs, and a user study shows that our results are preferred over other techniques.

We present an efficient technique to optimize color consistency of a collection of images depicting a common scene. Our method first recovers sparse pixel correspondences in the input images and stacks them into a matrix with many missing entries. We show that this matrix satisfies a rank two constraint under a simple color correction model. These parameters can be viewed as pseudo white balance and gamma correction parameters for each input image. We present a robust low-rank matrix factorization method to estimate the unknown parameters of this model. Using them, we improve color consistency of the input images or perform color transfer with any input image as the source. Our approach is insensitive to outliers in the pixel correspondences thereby precluding the need for complex pre-processing steps. We demonstrate high quality color consistency results on large photo collections of popular tourist landmarks and personal photo collections containing images of people.

This paper presents a vision system and a depth processing algorithm for DRC-HUBO+, the winner of the DRC finals 2015. Our system is designed to reliably capture 3D information of a scene and objects and to be robust to challenging environment conditions. We also propose a depth-map upsampling method that produces an outliers-free depth map by explicitly handling depth outliers. Our system is suitable for robotic applications in which a robot interacts with the real-world, requiring accurate object detection and pose estimation. We evaluate our depth processing algorithm in comparison with state-of-the-art algorithms on several synthetic and real-world datasets.

With the advent of commodity autonomous mobiles, it is becoming increasingly prevalent to recognize under extreme conditions such as night, erratic illumination conditions. This need has caused the approaches using multi-modal sensors, which could be complementary to each other. The choice for the thermal camera provides a rich source of temperature information, less affected by changing illumination or background clutters. However, existing thermal cameras have a relatively smaller resolution than RGB cameras that has trouble for fully utilizing the information in recognition tasks. To mitigate this, we aim to enhance the low-resolution thermal image according to the extensive analysis of existing approaches. To this end, we introduce Thermal Image Enhancement using Convolutional Neural Network (CNN), called in TEN, which directly learns an end-to-end mapping a single low resolution image to the desired high resolution image. In addition, we examine various image domains to find the best representative of the thermal enhancement. Overall, we propose the first thermal image enhancement method based on CNN guided on RGB data. We provide extensive experiments designed to evaluate the quality of image and the performance of several object recognition tasks such as pedestrian detection, visual odometry, and image registration.

Autonomous driving technology has made significant advances in recent years. In order for self-driving cars to become practical, they are required to operate safely and reliably even under adverse driving conditions. However, most current autonomous driving cars have only been shown to be operational under amiable weather conditions, i.e., on sunny days on dry roads. In order to enable autonomous cars to handle adverse driving conditions such as rain and wet roads, the algorithm must be able to detect roads within a tolerable margin of error using sensors such as cameras and laser scanners. In this paper, we propose a sensor fusion algorithms that is able to operate under a variety of weather conditions, including rain. Our algorithm was validated when a strong shower occurred during the 2014 Hyundai Motor Company's Autonomous Car Competition. In this paper, we present the competition results that were collected on the same course on both sunny and rainy days. Based on the comparison, we propose the future directions to improve the autonomous driving capability under adverse environmental conditions.

We present an object proposal method which utilizes the 3D data obtained from a depth sensor as well as the color information of images. Our object proposal method is designed to improve the performance of the object detection for a mobile robot equipped with a camera and a laser scanner. Compared to traditional object proposal methods using only 2D images, the proposed method provides much less number of candidate windows for object detection. We show less than 100 object proposal windows per image using the proposed method result in high recall tested on the public dataset. Our method presents object proposals in 3D space as well as in 2D image thus it can further be applied to following tasks for mobile robots such as 3D location and pose estimation of the target object after successful object detection. We validate our method using the real-world object detection dataset for outdoor mobile robots captured during the DRC Finals 2015 and the public dataset for comparison with the previous methods.

Drivable region detection is challenging since various types of road, occlusion or poor illumination condition have to be considered in a outdoor environment, particularly at night. In the past decade, Many efforts have been made to solve these problems, however, most of the already existing methods are designed for visible light cameras, which are inherently inefficient under low light conditions. In this paper, we present a drivable region detection algorithm designed for thermal-infrared cameras in order to overcome the aforementioned problems. The novelty of the proposed method lies in the utilization of on-line road initialization with a highly scene-adaptive sampling mask. Furthermore, our prior road information extraction is tailored to enforce temporal consistency among a series of images. In this paper, we also propose a large number of experiments in various scenarios (on-road, off-road and cluttered road). A total of about 6000 manually annotated images are made available in our website for the research community. Using this dataset, we compared our method against multiple state-of-the-art approaches including convolutional neural network (CNN) based methods to emphasize the robustness of our approach under challenging situations.

Commonly used in many applications, robust PCA represents an algorithmic attempt to reduce the sensitivity of classical PCA to outliers. The basic idea is to learn a decomposition of some data matrix of interest into low rank and sparse components, the latter representing unwanted outliers. Although the resulting problem is typically NP-hard, convex relaxations provide a computationally-expedient alternative with theoretical support. However, in practical regimes performance guarantees break down and a variety of non-convex alternatives, including Bayesian-inspired models, have been proposed to boost estimation quality. Unfortunately though, without additional a priori knowledge none of these methods can significantly expand the critical operational range such that exact principal subspace recovery is possible. Into this mix we propose a novel pseudo-Bayesian algorithm that explicitly compensates for design weaknesses in many existing non-convex approaches leading to state-of-the-art performance with a sound analytical foundation.

Robust Principal Component Analysis (RPCA) via rank minimization is a powerful tool for recovering underlying low-rank structure of clean data corrupted with sparse noise/outliers. In many low-level vision problems, not only it is known that the underlying structure of clean data is low-rank, but the exact rank of clean data is also known. Yet, when applying conventional rank minimization for those problems, the objective function is formulated in a way that does not fully utilize a priori target rank information about the problems. This observation motivates us to investigate whether there is a better alternative solution when using rank minimization. In this paper, instead of minimizing the nuclear norm, we propose to minimize the partial sum of singular values, which implicitly encourages the target rank constraint. Our experimental analyses show that, when the number of samples is deficient, our approach leads to a higher success rate than conventional rank minimization, while the solutions obtained by the two approaches are almost identical when the number of samples is more than sufficient. We apply our approach to various low-level vision problems, e.g., high dynamic range imaging, motion edge detection, photometric stereo, image alignment and recovery, and show that our results outperform those obtained by the conventional nuclear norm rank minimization method.

In this paper, we propose a classification approach for ambiguously shaped defects found on the surface of a type of display panel module that is widely used in the field of mobile displays. These types of surface defects are difficult to properly distinguish due to defect similarity and diversity. In such cases, defect types can only be determined using cumbersome human visual inspection. To solve the problem of ambiguous surface defect classification, we introduce a novel filtering method that effectively separates the foreground defective regions from the background, which has structured patterns, local illumination variation, and different light conditions for each of several cameras in an inspection system. Applying the proposed filter method to defect images, we select important features by adopting a wrapper-based feature selection method using a random forest as a learning algorithm. Successful classification results using the presented model are obtained using challenging real-world defect image data gathered from a smart phone display module inspection line in an industrial plant.

This paper presents an automatic method to extract a multi-view object in a natural environment. We assume that the target object is bounded by the convex volume of interest defined by the overlapping space of camera viewing frustums. There are two key contributions of our approach. First, we present an automatic method to identify a target object across different images for multi-view binary co-segmentation. The extracted target object shares the same geometric representation in space with a distinctive color and texture model from the background. Second, we present an algorithm to detect color ambiguous regions along the object boundary for matting refinement. Our matting region detection algorithm is based on information theory, which measures the Kullback-Leibler (KL) divergence of local color distribution of different pixel-bands. The local pixel-band with the largest entropy is selected for matte refinement, subject to the multi-view consistent constraint. Our results are highquality alpha mattes consistent across all different viewpoints. We demonstrate the effectiveness of the proposed method using various examples.

This paper describes a new methodology for estimating the relative pose between two 2-D lidars. Scanned points of 2-D lidars do not have enough feature information for correspondence matching. For this reason, additional image sensors or artificial landmarks at known locations have been used to find the relative pose. We propose a novel method of estimating the relative pose between 2-D lidars without any additional sensors or artificial landmarks. By scanning two orthogonal planes, we utilize the coplanarity of the scan points on each plane and the orthogonality of the plane normals. Even if we capture planes which are not exactly orthogonal, the method provides good results using nonlinear optimization. Experiments with both synthetic and real data show the validity of the proposed method. We also derive two degenerate cases: one related to plane poses, and the other caused by the relative pose. To the best of our knowledge, this study provides the first solution for the problem.

One of the most hazardous driving scenario is the overtaking of a slower vehicle, indeed, in this case the front vehicle (being overtaken) can occlude an important part of the field of view of the rear vehicle's driver. This lack of visibility is the most probable cause of accidents in this context. Recent research works tend to prove that augmented reality applied to assisted driving can significantly reduce the risk of accidents. In this paper, we present a real-time marker-less system to see through cars. For this purpose, two cars are equipped with cameras and an appropriate wireless communication system. The stereo vision system mounted on the front car allows to create a sparse 3D map of the environment where the rear car can be localized. Using this inter-car pose estimation, a synthetic image is generated to overcome the occlusion and to create a seamless see-through effect which preserves the structure of the scene.

We present a multi-view structured-light system which uses geometric and photometric information observed at multiple viewpoints. In our method, a highly accurate geometry reconstruction is achieved by exploiting our multi-view constraint on unwrapped phase images instead of using photo-consistency between intensity images. Also to express fine 3D information, we directly use photometric normal to infer non-linear shape between each point spacing, rather than fusing it by refining the geometry. Our system is built upon the off-the-shelf structured-light system, so that the proposed algorithm can be applied to existing systems including commercial 3D scanners to improve the accuracy without any system modification. We evaluate the performance of our method on both synthetic and real data, and demonstrate superior accuracy, which surpasses the accuracy of the state-of-the-art commercial products.

Single-shot structured light scanning has been actively investigated as it can recover accurate geometrical shape even on a dynamic scene. Since many single-shot approaches focus on improving depth accuracy, recovering the intrinsic properties of the scene such as albedo and shading are also valuable. In this paper, we propose a novel method that reconstructs not only the metric depth but also the intrinsic properties from a single structured light image. We extend the conventional color structured light model to embrace the Lambertian shading model. By using a color phase-shifting pattern, we parameterize the captured image with only two variables, albedo and depth. For an initial solution, a simple but powerful method to decompose sinusoids from the input image is presented. We formulate a non-linear cost function and jointly optimize albedo and depth efficiently by calculating analytic Jacobian. We demonstrate that our algorithm reasonably works on various real-world objects which exhibit challenging surface reflectance and albedo.

With the increasing interest in pedestrian detection, pedestrian datasets have also been the subject of research in the past decades. However, most existing datasets focus on a color channel, while a thermal channel is helpful for detection even in a dark environment. With this in mind, we propose a multispectral pedestrian dataset which provides well aligned color-thermal image pairs, captured by beam splitter-based special hardware. The color-thermal dataset is as large as previous color-based datasets and provides dense annotations including temporal correspondences. With this dataset, we introduce multispectral ACF, which is an extension of aggregated channel features (ACF) to simultaneously handle color-thermal image pairs. Multi-spectral ACF reduces the average miss rate of ACF by 15%, and achieves another breakthrough in the pedestrian detection task.

This paper introduces an algorithm that accurately estimates depth maps using a lenslet light field camera. The proposed algorithm estimates the multi-view stereo correspondences with sub-pixel accuracy using the cost volume. The foundation for constructing accurate costs is threefold. First, the sub-aperture images are displaced using the phase shift theorem. Second, the gradient costs are adaptively aggregated using the angular coordinates of the light field. Third, the feature correspondences between the sub-aperture images are used as additional constraints. With the cost volume, the multi-label optimization propagates and corrects the depth map in the weak texture regions. Finally, the local depth map is iteratively refined through fitting the local quadratic function to estimate a non-discrete depth map. Because micro-lens images contain unexpected distortions, a method is also proposed that corrects this error. The effectiveness of the proposed algorithm is demonstrated through challenging real world examples and including comparisons with the performance of advanced depth estimation algorithms.

We present an outdoor photometric stereo method using images captured in a single day. We simulate a sky hemisphere for each image according to its GPS and timestamp, and parameterize the obtained sky hemisphere into a quadratic skylight and a Gaussian sunlight distribution. Unlike previous works which usually model outdoor illumination as a sum of constant ambient light and a distant point light, our method models natural illumination according to a popular sky model and thus provides sufficient constraints for shape reconstruction from one day images. We generate pixel profiles of uniformly sampled unit vectors for the corresponding time of captures and evaluate them using correlation with the actual pixel profiles. The estimated surface normal is refined by MRF optimization. We have tested our method to recover objects and scenes of various sizes in real-world outdoor daylight.

Rank minimization problem can be boiled down to either Nuclear Norm Minimization (NNM) or Weighted NNM (WNNM) problem. The problems related to NNM (or WNNM) can be solved iteratively by applying a closed-form proximal operator, called Singular Value Thresholding (SVT) (or Weighted SVT), but they suffer from high computational cost to compute a Singular Value Decomposition (SVD) at each iteration. In this paper, we propose an accurate and fast approximation method for SVT, called fast randomized SVT (FRSVT), where we avoid direct computation of SVD. The key idea is to extract an approximate basis for the range of a matrix from its compressed matrix. Given the basis, we compute the partial singular values of the original matrix from a small factored matrix. While the basis approximation is the bottleneck, our method is already severalfold faster than thin SVD. By adopting a range propagation technique, we can further avoid one of the bottleneck at each iteration. Our theoretical analysis provides a stepping stone between the approximation bound of SVD and its effect to NNM via SVT. Along with the analysis, our empirical results on both quantitative and qualitative studies show our approximation rarely harms the convergence behavior of the host algorithms. We apply it and validate the efficiency of our method on various vision problems, e.g. subspace clustering, weather artifact removal, simultaneous multi-image alignment and rectification.

We propose a novel camera calibration method for defocused images using a smartphone under the assumption that the defocus blur is modeled as a convolution of a sharp image with a Gaussian point spread function (PSF). In contrast to existing calibration approaches which require well-focused images, the proposed method achieves accurate camera calibration with severely defocused images. This robustness to defocus is due to the proposed set of unidirectional binary patterns, which simplifies 2D Gaussian deconvolution to a 1D Gaussian deconvolution problem with multiple observations. By capturing the set of patterns consecutively displayed on a smartphone, we formulate the feature extraction as a deconvolution problem to estimate feature point locations in sub-pixel accuracy and the blur kernel in each location. We also compensate the error in camera parameters due to refraction of the glass panel of the display device. We evaluate the performance of the proposed method on synthetic and real data. Even under severe defocus, our method shows accurate camera calibration result.

We present a practical 3D reconstruction method to obtain a high-quality dense depth map from narrow-baseline image sequences captured by commercial digital cameras, such as DSLRs or mobile phones. Depth estimation from small motion has gained interest as a means of various photographic editing, but important limitations present themselves in the form of depth uncertainty due to a narrow baseline and rolling shutter. To address these problems, we introduce a novel 3D reconstruction method from narrow-baseline image sequences that effectively handles the effects of a rolling shutter that occur from most of commercial digital cameras. Additionally, we present a depth propagation method to fill in the holes associated with the unknown pixels based on our novel geometric guidance model. Both qualitative and quantitative experimental results show that our new algorithm consistently generates better 3D depth maps than those by the state-of-the-art method.

In this paper, we present a novel multi-image motion deblurring method utilizing the coded exposure technique. The key idea of our work is to capture video frames with a set of complementary fluttering patterns to preserve spatial frequency details. We introduce an algorithm for generating a complementary set of binary sequences based on the modern communication theory and implement the coded exposure video system with an off-the-shelf machine vision camera. The effectiveness of our method is demonstrated on various challenging examples with quantitative and qualitative comparisons to other computational image capturing methods used for image deblurring.

We present a novel detection method using a deep convolutional neural network (CNN), named AttentionNet. We cast an object detection problem as an iterative classification problem, which is the most suitable form of a CNN. AttentionNet provides quantized weak directions pointing a target object and the ensemble of iterative predictions from AttentionNet converges to an accurate object boundary box. Since AttentionNet is a unified network for object detection, it detects objects without any separated models from the object proposal to the post bounding-box regression. We evaluate AttentionNet by a human detection task and achieve the state-of-the-art performance of 65% (AP) on PASCAL VOC 2007/2012 with an 8-layered architecture only.

We present a method to reconstruct dense 3D points from small camera motion. We begin with estimating sparse 3D points and camera poses by Structure from Motion (SfM) method with homography decomposition. Although the estimated points are optimized via bundle adjustment and gives reliable accuracy, the reconstructed points are sparse because it heavily depends on the extracted features of a scene. To handle this, we propose a depth propagation method using both a color prior from the images and a geometry prior from the initial points. The major benefit of our method is that we can easily handle the regions with similar colors but different depths by using the surface normal estimated from the initial points. We design our depth propagation framework into the cost minimization process. The cost function is linearly designed, which makes our optimization tractable. We demonstrate the effectiveness of our approach by comparing with a conventional method using various real-world examples.

We propose a spatially varying stitching method with line correspondences. We are motivated by the observation that point features could be spatially biased or not matched in practice, e.g., repeated textures or homogeneous regions of man-made structures. In this scenario, line matches can provide strong correspondences as well as supplement cues, such as the structure preserving property. With these advantages, we adopt a feature fusion method that combines point and line correspondences into a unified framework for spatially varying stitching. We then estimate the balancing parameter between the point and line terms using geometric error. Our experiments show accurate alignment for challenging but common cases.

As people are becoming interested in paintings, various user-interactive search systems have been presented in recent times. Many systems encourage users to search paintings by prior knowledge on paintings. We discover the limitation for existing methods on how well the query is represented by the user, and propose a simple, yet effective way to search the painting by exploiting the color to express human visual memory. To achieve our goal, we suggest color clustering based on human color perception, and hierarchical metric learning to accommodate the locality of colors. With user-interactive drawing through learned colors, the user completes the abstract image to resemble the visual memory. We show that our system is easy to use, fast to process, accurate to search and fully extensible to cover deviation among users.

The purpose of this paper is to introduce a new method for removing reflections from multi-view images taken through a transparent medium, such as pane glass. Our method utilizes an optimization approach based on the probabilistic model of relative smoothness algorithm, which exploits gradient value to separate the image into two sub-layers. As this algorithm has certain limitations on removing reflections, we improve upon it by imposing a gradient-sparsity constraint. This allows the type of reflection captured within a camera's focal length to be effectively removed. We also introduce another constraint on a disparity map which smooths specific areas of reflection layer while simultaneously preserves the sharpness of the main object. These two major contributions are proven to be sufficient in producing high-quality images. Our algorithm demonstrates good results compared with other existing methods as most of the reflection spots have been removed, and the computational time of our system is arguably fast.

We present a calibration method of a time-of-flight (ToF) sensor and a color camera pair to align the 3D measurements with the color image correctly. We have designed a 2.5D pattern board with irregularly placed holes to be accurately detected from low resolution depth images of a ToF camera as well as from high resolution color images. In order to improve the accuracy of the 3D measurements of a ToF camera, we propose to perform ray correction and range bias correction. We reset the transformation of the ToF sensor which transforms the radial distance into the scene depth in Cartesian coordinate through ray correction. Then we capture a planar scene from different depths to correct the distance error that is shown to be dependent not only on the distance but also on the pixel location. The range error profiles along the calibrated distance are classified according to their wiggling shapes and each cluster of profiles with similar shape are separately estimated using a B-spline function. The standard deviation of the remaining random noise is recorded as an uncertainty information of distance measurements. We show the performance of our calibration method quantitatively and qualitatively on various datasets, and validate the impact of our method by demonstrating an RGB-D shape refinement application.

This paper introduces a new high dynamic range (HDR) imaging algorithm which utilizes rank minimization. Assuming a camera responses linearly to scene radiance, the input low dynamic range (LDR) images captured with different exposure time exhibit a linear dependency and form a rank-1 matrix when stacking intensity of each corresponding pixel together. In practice, misalignments caused by camera motion, presences of moving objects, saturations and image noise break the rank-1 structure of the LDR images. To address these problems, we present a rank minimization algorithm which simultaneously aligns LDR images and detects outliers for robust HDR generation. We evaluate the performances of our algorithm systematically using synthetic examples and qualitatively compare our results with results from the state-of-the-art HDR algorithms using challenging real world examples.

Recently, there have been significant advances in self-driving cars, which will play key roles in future intelligent transportation systems. In order for these cars to be successfully deployed on real roads, they must be able to autonomously drive along collision-free paths while obeying traffic laws. In contrast to many existing approaches that use prebuilt maps of roads and traffic signals, we propose algorithms and systems using Unified Map built with various onboard sensors to detect obstacles, other cars, traffic signs, and pedestrians. The proposed map contains not only the information on real obstacles nearby but also traffic signs and pedestrians as virtual obstacles. Using this map, the path planner can efficiently find paths free from collisions while obeying traffic laws. The proposed algorithms were implemented on a commercial vehicle and successfully validated in various environments, including the 2012 Hyundai Autonomous Ground Vehicle Competition.

In this paper, we propose a method to refine geometry of 3D meshes from the Kinect fusion by exploiting shading cues captured from the infrared (IR) camera of Kinect. A major benefit of using the Kinect IR camera instead of a RGB camera is that the IR images captured by Kinect are narrow band images which filtered out most undesired ambient light that makes our system robust to natural indoor illumination. We define a near light IR shading model which describes the captured intensity as a function of surface normals, albedo, lighting direction, and distance between a light source and surface points. To resolve ambiguity in our model between normals and distance, we utilize an initial 3D mesh from the Kinect fusion and multi-view information to reliably estimate surface details that were not reconstructed by the Kinect fusion. Our approach directly operates on a 3D mesh model for geometry refinement. The effectiveness of our approach is demonstrated through several challenging real-world examples.

This paper introduces a new color transfer method which is a process of transferring color of an image to match the color of another image of the same scene. The color of a scene may vary from image to image because the photographs are taken at different times, with different cameras, and under different camera settings. To solve for a full nonlinear and nonparametric color mapping in the 3D RGB color space, we propose a scattered point interpolation scheme using moving least squares and strengthen it with a probabilistic modeling of the color transfer in the 3D color space to deal with mis-alignments and noise. Experiments show the effectiveness of our method over previous color transfer methods both quantitatively and qualitatively. In addition, our framework can be applied for various instances of color transfer such as transferring color between different camera models, camera settings, and illumination conditions, as well as for video color transfers.

We show that a non-isotropic near point light source rigidly attached to a camera can be calibrated using multiple images of a weakly textured planar scene. We prove that if the radiant intensity distribution (RID) of a light source is radially symmetric with respect to its dominant direction, then the shading observed on a Lambertian scene plane is bilaterally symmetric with respect to a 2D line on the plane. The symmetry axis detected in an image provides a linear constraint for estimating the dominant light axis. The light position and RID parameters can then be estimated using a linear method. Specular highlights if available can also be used for light position estimation. We also extend our method to handle non-Lambertian reflectances which we model using a biquadratic BRDF. We have evaluated our method on synthetic data quantitavely. Our experiments on real scenes show that our method works well in practice and enables light calibration without the need of a specialized hardware.

A typical gaze estimator needs an explicit personal calibration stage with many discrete fixation points. This limitation can be resolved by mapping multiple eye images and corresponding saliency maps of a video clip during an implicit calibration stage. Compared to previous calibration-free methods, our approach clusters eye images by using Gaussian Mixture Model (GMM) in order to increase calibration accuracy and reduce training redundancy. Eye feature vectors representing eye images undergo soft clustering with GMM as well as the corresponding saliency maps for aggregation. The GMM based soft-clustering boosts the accuracy of Gaussian process regression which maps between eye feature vectors and gaze directions given this constructed data. The experimental results show an increase in gaze estimation accuracy compared to previous works on calibration-free method.

We present an efficient optical flow refinement approach based on a bidirectional flow consistency. Our method is an add-on component that improves the existing optical flow estimation to be balanced between forward and backward flows. Most of the state-of-the-art optical flow methods only consider unidirectional motion vectors from a source image to a target image, which can make the estimated flow inconsistent with its backward estimation. The inconsistency can be reduced by considering the bidirectional motion when the optical flow is estimated, but it would be very hard for most of the typical optical flow methods and impossible for some of them. To solve this problem, we propose a sampling-based optimization method for efficiently refining the optical flows with a bidirectional constraint. By evaluating on Middle-bury benchmark and public large displacement datasets, we validate the effectiveness of our method quantitatively and qualitatively and for accuracy.

Since commercial light field cameras became available, the light field camera has aroused much interest from computer vision and image processing communities due to its versatile functions. Most of its special features are based on an estimated depth map, so reliable depth estimation is a crucial step. However, estimating depth on real light field cameras is a challenging problem due to noise and short baselines among sub-aperture images. We propose a depth map estimation method for light field cameras by exploiting correspondence and focus cues. We aggregate costs among all the sub-aperture images on cost volume to alleviate noise effects. With efficiency of the cost volume, cost-aware depth estimation is quickly achieved by discrete-continuous optimization. In addition, we analyze each property of correspondence and focus cues and utilize them to select reliable anchor points. A well reconstructed initial depth map from the anchors is shown to enhance convergence. We show our method outperforms the state-of-the-art methods by validating it on real datasets acquired with a Lytro camera.

This paper describes a new methodology for estimating a relative pose of two 2D laser sensors. Two dimensional laser scan points do not have enough feature information for motion tracking. For this reason, additional image sensors or artificial landmarks have been used to find a relative pose. We propose the method to estimate a relative pose of 2D laser sensors without any additional sensor or artificial landmark. By scanning two orthogonal planes, we utilize only the coplanarity of the scan points on each plane and the orthogonality of the plane normals. Experiments with both synthetic and real data show the validity of the proposed method. To the best of our knowledge this works provides the first solution for the problem.

In this paper we propose a hybrid vision-based SLAM and moving objects tracking (vSLAMMOT) approach. This approach tightly combines two key methods: a superpixel-based segmentation to detect moving objects and a Rao-Blackwellized Particle Filter to estimate a stereo-vision-based SLAM posterior. Most successful methods perform vision-based SLAM (vSLAM) and track moving objects independently. However, we pose both vSLAM and moving object tracking as a single correlated problem to leverage the performance. Our approach estimates the relative camera motion using the previous tracking result, and then detects moving objects from the estimated camera motion recursively. Moving superpixels are detected by a Markov Random Field (MRF) model which uses spatial and temporal information of the moving objects. We demonstrate the performance of the proposed approach for vSLAMMOT using both synthetic and real datasets and compare the performance with other methods.

In this paper are presented simple and practical solutions to extrinsic calibration between a camera and a 2D laser sensor, without overlap. Previous methods utilized a plane or an intersecting line of two planes as a geometric constraint with enough common field-of-view. These required additional sensors to calibrate non-overlapping systems. In this paper, we present two methods for solving the problem - one utilizes a plane; the other utilizes an intersecting line of two planes. For each method, an initial solution of the relative positions of a non-overlapping camera and a laser sensor, was computed by adopting a reasonable assumption about geometric structures. Then we refined it via non-linear optimization, even if the assumption was not perfectly satisfied. Both simulation results and experiments using real data showed that the proposed methods provided reliable results compared to ground-truth, and similar or better results than those provided by a conventional method.

Accurate estimation of camera motion is very important for many robotics applications involving SfM and visual SLAM. Such accuracy is attempted by refining the estimated motion through nonlinear optimization. As many modern robots are equipped with both 2D and 3D cameras, it is both highly desirable and challenging to exploit data acquired from both modalities to achieve a better localization. Existing refinement methods, such as Bundle adjustment and loop closing, may be employed only when precise 2D-to-3D correspondences across frames are available. In this paper, we propose a framework for robot localization that benefits from both 2D and 3D information without requiring such accurate correspondences to be established. This is carried out through a 2D-3D based initial motion estimation followed by a constrained nonlinear optimization for motion refinement. The initial motion estimation finds the best possible 2D-to-3D correspondences and localizes the cameras with respect the 3D scene. The refinement step minimizes the projection errors of 3D points while preserving the existing relationships between images. The problems of occlusion and that of missing scene parts are handled by comparing the image-based reconstruction and 3D sensor measurements. The effect of data inaccuracies is minimized using an M-estimator based technique. Our experiments have demonstrated that the proposed framework allows to obtain a good initial motion estimate and a significant improvement through refinement.

We present a new method to auto-adjust camera exposure for outdoor robotics. In outdoor environments, scene dynamic range may be wider than the dynamic range of the cameras due to sunlight and skylight. This can results in failures of vision-based algorithms because important image features are missing due to under-/over-saturation. To solve the problem, we adjust camera exposure to maximize image features in the gradient domain. By exploiting the gradient domain, our method naturally determines the proper exposure needed to capture important image features in a manner that is robust against illumination conditions. The proposed method is implemented using an off-the-shelf machine vision camera and is evaluated using outdoor robotics applications. Experimental results demonstrate the effectiveness of our method, which improves the performance of robot vision algorithms.

In real-time memory-intensive image processing and vision applications, increasing image resolution requires the use of external SDR/DDR memories. However, the arbitrary pixel access patterns used in most algorithms reduce their memory throughput as a result of increasing access latency. Efficient cache design is paramount in real-time memory-intensive applications. Its effectiveness depends on the spatial and temporal locality of data access. In image processing, the spatial locality denotes the neighboring pixels, located horizontally and vertically in 2-D. However, the conventional caches used in general processors cannot define the vertical locality. We propose a rolling cache optimized for image format and algorithms, a method to reduce the miss penalty by moving the cache horizontally and vertically, and a parallel processing architecture with interpolation, multilevel and multiple caches. To support our idea, we compare it with other types of caches and show that the average memory access time and the memory bandwidth are decreased by 28% and 74%, respectively, for a 2048 × 2048 image. Its performance is greater than that of the 16-way set associative cache, but the tag memory size is a bit larger than that of the direct-mapped cache. Using two different applications, we show that the proposed architecture is applicable to a number of algorithms if data access follows an arbitrary curve or block-wise pattern, which is the usual case with image processing and vision algorithms. If an application is based on local data access in resource-limited systems, it is possible to achieve high performance with lower operational frequency using the proposed architecture.

This paper describes an application framework to perform high-quality upsampling and completion on noisy depth maps. Our framework targets a complementary system setup, which consists of a depth camera coupled with an RGB camera. Inspired by a recent work that uses a nonlocal structure regularization, we regularize depth maps in order to maintain fine details and structures. We extend this regularization by combining the additional high-resolution RGB input when upsampling a low-resolution depth map together with a weighting scheme that favors structure details. Our technique is also able to repair large holes in a depth map with consideration of structures and discontinuities utilizing edge information from the RGB input. Quantitative and qualitative results show that our method outperforms existing approaches for depth map upsampling and completion. We describe the complete process for this system, including device calibration, scene warping for input alignment, and even how our framework can be extended for video depth-map completion with the consideration of temporal coherence.

We present a novel method to separate specular reflection from a single image. Separating an image into diffuse and specular components is an ill-posed problem due to lack of observations. Existing methods rely on a specular-free image to detect and estimate specularity, which however may confuse diffuse pixels with the same hue but a different saturation value as specular pixels. Our method is based on a novel observation that for most natural images the dark channel can provide an approximate specular-free image. We also propose a maximum a posteriori formulation which robustly recovers the specular reflection and chromaticity despite of the hue-saturation ambiguity. We demonstrate the effectiveness of the proposed algorithm on real and synthetic examples. Experimental results show that our method significantly outperforms the state-of-the-art methods in separating specular reflection.

Object detection in real images or videos is challenging because the shapes and sizes of objects vary significantly according to their poses, camera viewing direction, and partial occlusion. Previous detection methods employ sliding-window-based schemes that scan windows across an image, requiring many differently shaped windows to capture shape and size variation. In order to solve this problem, we propose an object detection method using hierarchical graph-based segmentation: color-consistent parts are obtained by part-level segmentation and category-consistent regions are found using object-level segmentation. Thus we can avoid scanning a lot of windows across whole images by using part-level segmentation and robustly detect the objects of various shapes and sizes by using object-level segmentation. In addition, we evaluate detection performance using various classifiers with our detection approach.

We present a novel framework to estimate detailed shape of diffuse objects with uniform albedo from a single RGB-D image. To estimate accurate lighting in natural illumination environment, we introduce a general lighting model consisting of two components: global and local models. The global lighting model is estimated from the RGB-D input using the low-dimensional characteristic of a diffuse reflectance model. The local lighting model represents spatially varying illumination and it is estimated by using the smoothly-varying characteristic of illumination. With both the global and local lighting model, we can estimate complex lighting variations in uncontrolled natural illumination conditions accurately. For high quality shape capture, a shape-from-shading approach is applied with the estimated lighting model. Since the entire process is done with a single RGB-D input, our method is capable of capturing the high quality shape details of a dynamic object under natural illumination. Experimental results demonstrate the feasibility and effectiveness of our method that dramatically improves shape details of the rough depth input.

Finding a good binary sequence is critical in determining the performance of the coded exposure imaging, but previous methods mostly rely on a random search for finding the binary codes, which could easily fail to find good long sequences due to the exponentially growing search space. In this paper, we present a new computationally efficient algorithm for generating the binary sequence, which is especially well suited for longer sequences. We show that the concept of the low autocorrelation binary sequence that has been well exploited in the information theory community can be applied for generating the fluttering patterns of the shutter, propose a new measure of a good binary sequence, and present a new algorithm by modifying the Legendre sequence for the coded exposure imaging. Experiments using both synthetic and real data show that our new algorithm consistently generates better binary sequences for the coded exposure problem, yielding better deblurring and resolution enhancement results compared to the previous methods for generating the binary codes.

Robust Principal Component Analysis (RPCA) via rank minimization is a powerful tool for recovering underlying low-rank structure of clean data corrupted with sparse noise/outliers. In many low-level vision problems, not only it is known that the underlying structure of clean data is low-rank, but the exact rank of clean data is also known. Yet, when applying conventional rank minimization for those problems, the objective function is formulated in a way that does not fully utilize a priori target rank information about the problems. This observation motivates us to investigate whether there is a better alternative solution when using rank minimization. In this paper, instead of minimizing the nuclear norm, we propose to minimize the partial sum of singular values. The proposed objective function implicitly encourages the target rank constraint in rank minimization. Our experimental analyses show that our approach performs better than conventional rank minimization when the number of samples is deficient, while the solutions obtained by the two approaches are almost identical when the number of samples is more than sufficient. We apply our approach to various low-level vision problems, e.g. high dynamic range imaging, photometric stereo and image alignment, and show that our results outperform those obtained by the conventional nuclear norm rank minimization method.

We propose a method for accurate 3D shape reconstruction using uncalibrated multiview photometric stereo. A coarse mesh reconstructed using multiview stereo is first parameterized using a planar mesh parameterization technique. Subsequently, multiview photometric stereo is performed in the 2D parameter domain of the mesh, where all geometric and photometric cues from multiple images can be treated uniformly. Unlike traditional methods, there is no need for merging view-dependent surface normal maps. Our key contribution is a new photometric stereo based mesh refinement technique that can efficiently reconstruct meshes with extremely fine geometric details by directly estimating a displacement texture map in the 2D parameter domain. We demonstrate that intricate surface geometry can be reconstructed using several challenging datasets containing surfaces with specular reflections, multiple albedos and complex topologies.

We present a method based on a hierarchical clustering to restore the 3D lines of structured environments. In previous approaches, the restoration of noisy 3D lines is a challenging problem because it is difficult to define a suitable similarity measure discriminative to other lines. Our motivation to overcome the difficulty is that most structured scenes consist of sets of parallel 3D lines with the same angular proximity, which provides a hierarchical similarity measure for structured 3D lines. Accordingly, our restoration method works in a manner that clustering is hierarchically performed on angular and distance levels. The 3D line restoration is then achieved by finding the center of each cluster. The framework also makes the clustered 3D lines align along the associated angular directions. We compare the proposed algorithm with methods using no knowledge of the angular information, and demonstrate its effectiveness through real-world experiments.

We present a high dynamic range (HDR) imaging algorithm that utilizes a modern rank minimization framework. Linear dependency exists among low dynamic range (LDR) images. However, global or local misalignment by camera motion and moving objects breaks down the low-rank structure of LDR images. The proposed algorithm simultaneously estimates global geometric transforms to align LDR images and detects moving objects and under-/over-exposed regions using a rank minimization approach. In the HDR composition step, structural consistency weighting is proposed to generate an artifact-free HDR image from an user-selected reference image. We demonstrate the robustness and effectiveness of the proposed method with real datasets.

This paper presents a new structure-from-motion (SFM) technique called `generalized laser three-point' algorithm. It is designed to estimate the motion of the camera-laser fusion system which consists of a 2D laser sensor and multiple cameras. The laser points are projected onto the images and tracked to other frames to be used as 3D-2D correspondences. However, the typical three-point algorithms cannot estimate the motion of the system if three points are collinear. Using the laser points as 3D points, this case happens frequently if the laser sensor scans a large plane (e.g. open ground). Even in that case, two frames of the laser data are not collinear if the system is moved while it captures the frames. Among three point correspondences required to estimate the motion, we select two points and the other point from different frames. We estimate the relative pose between the frames by solving an 8-degree polynomial equation. The experimental results show that the proposed algorithm is more appropriate for our fusion system than the previous algorithms.

In this paper, we present a particle-filter-based localization framework with decoupled visual measurements (image features) for process and measurement models. Thus our approach enables using camera-based motion estimation while achieving the independence between the process noise and the measurement noise in the Bayesian filtering framework. In addition, we alternately perform sequential and global localization on the basis of the marginal likelihood in order to avoid severe errors caused by incorrect data association.

Data correspondence/grouping under an unknown parametric model is a fundamental topic in computer vision. Finding feature correspondences between two images is probably the most popular application of this research field, and is the main motivation of our work. It is a key ingredient for a wide range of vision tasks, including three-dimensional reconstruction and object recognition. Existing feature correspondence methods are based on either local appearance similarity or global geometric consistency or a combination of both in some heuristic manner. None of these methods is fully satisfactory, especially in the presence of repetitive image textures or mismatches. In this paper, we present a new algorithm that combines the benefits of both appearance-based and geometry-based methods and mathematically guarantees a global optimization. Our algorithm accepts the two sets of features extracted from two images as input, and outputs the feature correspondences with the largest number of inliers, which verify both the appearance similarity and geometric constraints. Specifically, we formulate the problem as a mixed integer program and solve it efficiently by a series of linear programs via a branch-and-bound procedure. We subsequently generalize our framework in the context of data correspondence/grouping under an unknown parametric model and show it can be applied to certain classes of computer vision problems. Our algorithm has been validated successfully on synthesized data and challenging real images.

We present a robust radiometric calibration framework that capitalizes on the transform invariant low-rank structure in the various types of observations, such as sensor irradiances recorded from a static scene with different exposure times, or linear structure of irradiance color mixtures around edges. We show that various radiometric calibration problems can be treated in a principled framework that uses a rank minimization approach. This framework provides a principled way of solving radiometric calibration problems in various settings. The proposed approach is evaluated using both simulation and real-world datasets and shows superior performance to previous approaches.

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.

We introduce a method to repair an image which has been stamped by an identigram or a watermark. Our method is based on the cross-channel correlation which assures the co-occurrence of image discontinuities and correlation of color distributions across different color channels of an image. Using blind source separation, we find the transformation of color space which separates the structures of identigram and that of the original image into two different individual color channels. To repair the image contents in the corrupted channel, we formulate the problem using Bayes' rule where the prior and the likelihood probabilities are defined based on the cross-channel correlation assumption. We compare our results with results from inpainting and texture synthesis-based hole filling techniques. Our results are pleasable for real-world examples and have the maximum PSNR for synthetic examples.

We present a motion detection algorithm by a change detection filter matrix derived from Discrete Cosine Transform. Recently, a Fourier reconstruction scheme shows good results for motion detection. However, its computational cost is a major drawback. We revisit the problem and achieve two orders of magnitude faster than the previous algorithm with better performance. The proposed algorithm runs at about 800 frames per second for VGA resolution images on a consumer hardware by using only integer matrix multiplication and the symmetric property of the change detection filter matrix. In addition, our algorithm is fundamentally robust to sudden illumination changes because it works based on edge information. We verify our algorithm with challenging datasets that contain strong and sudden illumination changes.

We introduce a learning based photo composition model and its application on photo re-arrangement. In contrast to previous approaches which evaluate quality of photo composition using the rule of thirds or the golden ratio, we train a normalized saliency map from visually pleasurable photos taken by professional photographers. We use Principal Component Analysis (PCA) to analyze training data and build a Gaussian mixture model (GMM) to describe the photo composition model. Our experimental results show that our approach is reliable and our trained photo composition model can be used to improve photo quality through photo re-arrangement.

In this paper, we propose a Markov Chain Monte Carlo (MCMC) sampling method with the data-driven proposal distribution for six-degree-of-freedom (6-DoF) SLAM. Recently, visual odometry priors have been widely used as the process model in the SLAM formulation to improve the SLAM performance. However, modeling the uncertainties of incremental motions estimated by visual odometry is especially difficult under challenging conditions, such as erratic motion. For a particle-based model representation, it can represent the uncertainty of the camera motion well under erratic motion compared to the constant velocity model or a Gaussian noise model, but the manner of representing the proposal distribution and sampling the particles is extremely important, as we can maintain only a limited number of particles in the high-dimensional state space. Hence, we propose an effective sampling approach by exploiting MCMC sampling and the data-driven proposal distribution to propagate the particles. We demonstrate the performance of the proposed approach for 6-DoF SLAM using both synthetic and real datasets and compare the performance with those of other sampling methods.

Building maps of unknown environments is a critical factor for autonomous navigation and homing, and this problem is especially challenging in large-scale environments. Recently, sensor fusion systems such as combinations of cameras and laser sensors have become popular in the effort to ensure a general level of performance in this task. In this paper, we present a new homing method in a large-scale environment using a laser-camera fusion system. Instead of fusing data to form a single map builder, we adaptively select sensor data to handle environments which contain ambiguity. For autonomous homing, we propose a new mapping strategy for building a hybrid map and a return strategy for selecting the next target waypoints efficiently. The experimental results demonstrate that the proposed algorithm enables the autonomous homing of a robot in a large-scale indoor environments in real time.

By the laws of quantum physics, pixel intensity does not have a true value, but should be a random variable. Contrary to the conventional assumptions, the distribution of intensity may not be an additive Gaussian. We propose to directly model the intensity difference and show its validity by an experimental comparison to the conventional additive model. As a model of the intensity difference, we present a Skellam distribution derived from the Poisson photon noise model. This modeling induces a linear relationship between intensity and Skellam parameters, while conventional variance computation methods do not yield any significant relationship between these parameters under natural illumination. The intensity-Skellam line is invariant to scene, illumination, and even most of camera parameters. We also propose practical methods to obtain the line using a color pattern and an arbitrary image under natural illumination. Because the Skellam parameters that can be obtained from this linearity determine a noise distribution for each intensity value, we can statistically determine whether any intensity difference is caused by an underlying signal difference or by noise. We demonstrate the effectiveness of this new noise model by applying it to practical applications of background subtraction and edge detection.

In this paper, we present results and experiments with several methods for bundle adjustment, producing the fastest bundle adjuster ever published in terms of computation and convergence. From a computational perspective, the fastest methods naturally handle the block-sparse pattern that arises in a reduced camera system. Adapting to the naturally arising block-sparsity allows the use of BLAS3, efficient memory handling, fast variable ordering, and customized sparse solving, all simultaneously. We present two methods; one uses exact minimum degree ordering and block-based LDL solving and the other uses block-based preconditioned conjugate gradients. Both methods are performed on the reduced camera system. We show experimentally that the adaptation to the natural block sparsity allows both of these methods to perform better than previous methods. Further improvements in convergence speed are achieved by the novel use of embedded point iterations. Embedded point iterations take place inside each camera update step, yielding a greater cost decrease from each camera update step and, consequently, a lower minimum. This is especially true for points projecting far out on the flatter region of the robustifier. Intensive analyses from various angles demonstrate the improved performance of the presented bundler.

One fundamental problem of object tracking is the convergence of estimates to local maxima not corresponding to target objects. To mitigate this problem, constructing a good posterior distribution of the target state is important. In this letter, we propose a robust tracking approach by building a new posterior distribution model from multiple independent estimates of a target state. For each candidate of the target state, we compute a confidence score based on its spatial consistency with other estimates and photometric similarities with target models. Our posterior distribution model reflects tracking uncertainties well and adaptively defines the search region for the next frame. We validate the robustness of our approach on a number of challenging datasets.

We present a robust radiometric calibration method that capitalizes on the transform invariant low-rank structure of sensor irradiances recorded from a static scene with different exposure times. We formulate the radiometric calibration problem as a rank minimization problem. Unlike previous approaches, our method naturally avoids over-fitting problem; therefore, it is robust against biased distribution of the input data, which is common in practice. When the exposure times are completely unknown, the proposed method can robustly estimate the response function up to an exponential ambiguity. The method is evaluated using both simulation and real-world datasets and shows a superior performance than previous approaches.

One of the most important problems in visual tracking is how to incrementally update the appearance model because the appearance of a target object can be easily changed with time when the target is a deformable object or it is moving under varying illumination conditions. To solve these problems, we present a Rao-Blackwellized particle filter (RBPF)-based object tracking algorithm with the adaptive appearance model represented by a Gaussian mixture model (or a mixture of Gaussians model) because a single Gaussian reveals limita tions in modeling the target appearance when observations are corrupted by occlusion or the tracking error. We demonstrate the robustness of the proposed method using well-known databases, such as the CAVIAR and the PETS databases.

This paper describes an application framework to perform high quality upsampling on depth maps captured from a low-resolution and noisy 3D time-of-flight (3D-ToF) camera that has been coupled with a high-resolution RGB camera. Our framework is inspired by recent work that uses nonlocal means filtering to regularize depth maps in order to maintain fine detail and structure. Our framework extends this regularization with an additional edge weighting scheme based on several image features based on the additional high-resolution RGB input. Quantitative and qualitative results show that our method outperforms existing approaches for 3D-ToF upsampling. We describe the complete process for this system, including device calibration, scene warping for input alignment, and even how the results can be further processed using simple user markup.

We propose a novel and simple framework that solves two popular problems in digital photography: 2D face synthesis and 3D face modeling. 2D face synthesis aims at creating a new face, usually by mixing two or more portraits. We extend this notion to the combination of human and statue faces. The goal of 3D face modeling is to reconstruct a face in three dimensions from one or several images. These two tasks are often treated as separate problems although they both consider face modeling. In this paper, we propose a unified and general framework for both 2D and 3D cases that runs in a fully automatic manner. Our work also creates stereoscopic views for entertainment 3D display. Experimental results and subjective tests have confirmed the validity of our approach.

In this paper, we propose a graph-based shape matching method for deformable objects. In our approach, a graph is generated from an over-segmented input image, and the shape matching problem is treated as finding an optimal cycle in the graph. Given a shape template and a graph generated from the input, a product graph is generated to consider every possible correspondence between graph edges and template sub-parts. Because the proposed approach can estimate reasonable correspondences between a target object and a template, it is possible to extract the target object robustly in the presence of shape deformation and background clutter. The experiments on various examples are also presented to verify the performance of proposed method.

In this paper, we propose an automatic method to extract a foreground object captured from multiple viewpoints. We consider the foreground object is within the visual hull of camera field of views. By exploring the multi-view geometric relationship and color measurements of the input images, we can estimate the foreground segmentations as well as their fractional boundaries. To facilitate efficient computation and high quality mattes, we adopt a two-phase approach. The first phase of our algorithm provides quick and rough binary segmentations of the foreground object using graph-cut; the second phase refines the segmentation boundaries using matting. Our result is the high quality alpha mattes of the foreground object consistently across all different viewpoints. We demonstrate the effectiveness of our method using challenging examples.

In this paper, we present an accurate and robust 6D SLAM method that uses multiple 2D sensors, i.e. perspective cameras and planar laser scanners. We have investigated strengths and weaknesses of those two sensors for 6D SLAM by conducting specifically designed experiments, and found that the sensors can complement each other. In order to take full advantages of each approach, we fuse correspondences of those two sensors, rather than individually estimated motions. Correspondences obtained by the two sensors have different characteristics, but can be expressed in a common 2D-3D relation form. We use the correspondences in a single structure from-motion framework. In the initial motion estimation step, we propose a RANSAC-based method to generate and test multiple motion hypotheses by using multiple pools of correspondences, aiming to avoid potential bias of each sensor data. In the later motion refinement step, we introduce a variant of bundle adjustment to consider different types of constraints from the two sensors. The performance of the proposed method is demonstrated both quantitatively by experiments on closed loop sequences and qualitatively by large-scale experiments with DGPS trajectory. The proposed method successfully closes a loop of 320 meters in twenty thousand frames by incremental process only.

In this paper, we present a sensor fusion system of cameras and 2D laser sensors for 3D reconstruction. The proposed system is designed to capture data on a fast-moving ground vehicle. The system consists of six cameras and one 2D laser sensor. In order to capture data at high speed, we synchronized all sensors by detecting the laser ray at a specific angle and generating a trigger signal for the cameras. Reconstruction of 3D structures is done by estimating frame-by-frame motion and accumulating vertical laser scans. The difference between the proposed system and the previous works using two 2D laser sensors is that we do not assume 2D motion. The motion of the system in 3D space (including absolute scale) is estimated accurately by data-level fusion of images and range data. The problem of error accumulation is solved by loop closing, not by GPS. The moving objects are detected by utilizing the depth information provided by the laser sensor. The experimental results show that the estimated path is successfully overlayed on the satellite images.

Recently, many researchers have made efforts for accurate calibration of a Time-of-Flight camera to fully utilize its provided depth values. Yet most previous works focus mainly on intrinsic calibration by modeling its systematic errors and noises while extrinsic calibration is also an important factor when constructing sensor fusion system. In this paper, we present a calibration process that can correctly transfer the depth measurements onto the color image. We use 2.5D pattern so that sufficient reprojection error can be considered for both color and ToF cameras. The issues on obtaining the correct correspondences for this pattern are discussed. In the optimization stage, the depth constraint is also employed to ensure the depth measurements to lie on the pattern plane. The strengths of the proposed method over previous approaches are evaluated in several robotic applications which require precise ToF and camera calibration.

This paper addresses the problem of establishing correspondences between two sets of visual features using higher order constraints instead of the unary or pairwise ones used in classical methods. Concretely, the corresponding hypergraph matching problem is formulated as the maximization of a multilinear objective function over all permutations of the features. This function is defined by a tensor representing the affinity between feature tuples. It is maximized using a generalization of spectral techniques where a relaxed problem is first solved by a multidimensional power method and the solution is then projected onto the closest assignment matrix. The proposed approach has been implemented, and it is compared to state-of-the-art algorithms on both synthetic and real data.

Majority of CMOS image sensors in consumer market utilize a rolling shutter to increase sensitivity. However, it causes severe distortions, such as jitter, wobble, or skew. Since most of these kinds of sensors are used in hand-held devices, the approach of undistorting and generating stabilized images is restricted to resource limited systems. It has also been one of the major challenges to have a mathematical representation of CMOS rolling effect depicting the practical scenario, while keeping accuracy and stability. We propose that a CMOS sensor can be modeled by a section-wise charge-coupled devices model which has multiple homographies and exploit the observation that rolling shutter mechanism gives close relationships among them. We present a CMOS seven-parameter model, and show video stabilization algorithm by the iterative parameter estimation technique. We address four issues while accelerating our stabilization algorithm within resource limited environment: accuracy, stability, computation time, and resource utilization. We developed cache based optimization techniques to meet the requirement of the memory bandwidth and computational time for the iterative parameter estimation and final output image interpolation, and also proposed the incremental form of the seven-parameter model to greatly reduce resource consumption while maintaining the same results as the previous. The validity and effectiveness of our approach is demonstrated by experiments for different types of camera motions. The cache based optimization technique can be used to accelerate other types of iterative vision algorithms that require repetitive memory access: feature tracking, motion estimation, motion compensation, various types of image distortion correction, and also image warping and scaling.

Bag-of-words (BOW), which represents an image by the histogram of local patches on the basis of a visual vocabulary, has attracted intensive attention in visual categorization due to its good performance and flexibility. Conventional BOW neglects the contextual relations between local patches due to its Naïve Bayesian assumption. However, it is well known that contextual relations play an important role for human beings to recognize visual categories from their local appearance. This paper proposes a novel contextual bag-of-words (CBOW) representation to model two kinds of typical contextual relations between local patches, i.e., a semantic conceptual relation and a spatial neighboring relation. To model the semantic conceptual relation, visual words are grouped on multiple semantic levels according to the similarity of class distribution induced by them, accordingly local patches are encoded and images are represented. To explore the spatial neighboring relation, an automatic term extraction technique is adopted to measure the confidence that neighboring visual words are relevant. Word groups with high relevance are used and their statistics are incorporated into the BOW representation. Classification is taken using the support vector machine with an efficient kernel to incorporate the relational information. The proposed approach is extensively evaluated on two kinds of visual categorization tasks, i.e., video event and scene categorization. Experimental results demonstrate the importance of contextual relations of local patches and the CBOW shows superior performance to conventional BOW.

Most research on image decomposition, e.g., image segmentation and image parsing, has predominantly focused on the low-level visual clues within a single image and neglected the contextual information across images. In this paper, we present a new perspective to image decomposition piloted by the multilabel context associated with each individual image. Observing that the contextual information (i.e., local label representations of the same label are similar while those from different labels are dissimilar) exists across images, we propose to perform image decomposition in a collective way and obtain an optimal representation for each label from a set of multilabeled images. We formulate the problem as an optimization problem which maximizes inter-label difference while minimizing the intra-label difference of the target label representations and propose two ways to solve this problem. Such a contextual image decomposition has a wide variety of applications, among which two exemplary ones-multilabel image annotation and label ranking, are presented and evaluated with different classification techniques. Extensive experiments on two benchmark datasets demonstrate promising results.

In this paper, we present results and experiments with several methods for bundle adjustment, producing the fastest bundle adjuster ever published. The fastest methods work with the well known reduced camera system and handle the block-sparse pattern arising in the reduced camera system in a natural way. Adapting to the naturally arising block-sparsity allows the use of BLAS3, efficient memory handling, fast variable ordering, and customized sparse solving all at the same time. We present two methods, one using exact minimum degree ordering and block-based LDL solving, and one using block-based preconditioned conjugate gradient, both on the reduced camera system. We show experimentally that the adaptation to the natural block sparsity allows both these methods to perform better than previous ones. Further speed improvements are achieved by the novel use of embedded point iterations. The embedded point iterations take place inside each camera update step, yielding a higher cost decrease from each camera update step. This is especially true for points projecting far out on the flatter region of the robustifier.

In this paper, we present a new object detection method using codebook and boosted Markov Chain Monte Carlo (MCMC) estimation. It is relatively well detected using adaboost and simple Haar-like features for small objects. However, the detection problem is more difficult when object size becomes larger (over 150 × 150) due to different surface markings and clutter. Codebook-based object representation and boosted MCMC method can detect large objects robustly. Experimental results validate convincing detection for large objects.

Vision-based robotic applications such as Simultaneous Localization and Mapping (SLAM), global localization, and autonomous navigation have suffered from problems related to dynamic environments involving moving objects and kidnapping. One of the possible solutions to these problems is to establish robust correspondences when obtaining images from static scenes. Therefore we propose an efficient technique for determining correspondences to recover the current camera pose; in the proposed method, the FAST corner detector and SIFT descriptors are combined because in many methods for vision-based robotic applications, corner features have been adopted since they enable fast computation and simplify the computation of the correspondences between consecutive images. However, to recover the pose of the camera after kidnapping or at an unknown initial position, a robust feature matching algorithm is required because the pose of a camera is unlikely to be the same as the poses in the database images. For this purpose, first, we determine some candidates for correspondences by combining corners with their multiple descriptors computed from previously defined scales, and then we select one of these candidates by optimizing the scale using a variant of the mean-shift algorithm. We apply the proposed matching algorithm to kidnapping and visual occlusion problems in autonomous navigation.

Particle filters have been used for visual tracking during long periods because they enable effective estimation for non-linear and non-Gaussian distributions. However, particle filter-based tracking approaches suffer from occlusion and deformation of the target objects, which result in the large difference between the current observations and the target model. Thus, we present a Rao-Blackwellized particle filter (RBPF)-based tracking algorithm that effectively estimates the joint distribution for the target state and the target model; in the proposed method, the target object is tracked by using the particle filter while the target model is simultaneously updated on the basis of the on-line approximation of a mixture of Gaussians. To ensure the robustness to occlusion, we represent the target model by 16 orientation histograms that are spatially divided, and individually update each histogram through a video sequence. We demonstrate the robustness of the proposed method under occlusion and deformation of the target objects.

Whereas some methods for plane extraction have been proposed, this problem still remains an open issue due to the complexity of the task. This paper especially focuses on the extraction of points lying on a plane (such as the ground and buildings walls) in sequences acquired by a central omnidirectional camera. Our approach is based on the epipolar constraint for planar scenes (i.e. homography) on a pair of omnidirectional images to detect some interest points belonging to a plane. Our main contribution is the introduction of a new method, called “2-point algorithm for homography”, that imposes some constraints on the homography using vanishing point (VP) information. Compared to the widely used DLT (4-point) algorithm, experiments on real data demonstrated that the proposed “2-point algorithm for homography” is more robust to noise and false matching, even when the plane to extract is not dominant in the image. Finally, we show that our system provides key clues for ground segmentation by GrabCut.

We present a method for simultaneously locking on to a ground target and estimating the position of an unmanned aerial vehicle (UAV) under countermeasure (CM) conditions, where sensors are prevented from successfully tracking a target. Owing to the limited payload and power of the UAVs, we employ a monocular camera and a global positioning system (GPS) to carry out vision-based simultaneous localization and mapping (SLAM) using both an unscented Kalman filter and a Kalman filter. Since this approach estimates the state of the UAV and the location of the target, we can estimate the position of the target in the image, even in the presence of CMs. Our experiments show that the proposed method successfully locks on to the target and estimates the state of the UAV.

International audience

This paper addresses the problem of establishing correspondences between two sets of visual features using higher-order constraints instead of the unary or pairwise ones used in classical methods. Concretely, the corresponding hypergraph matching problem is formulated as the maximization of a multilinear objective function over all permutations of the features. This function is defined by a tensor representing the affinity between feature tuples. It is maximized using a generalization of spectral techniques where a relaxed problem is first solved by a multi-dimensional power method, and the solution is then projected onto the closest assignment matrix. The proposed approach has been implemented, and it is compared to state-of-the-art algorithms on both synthetic and real data.

Most research on image decomposition, e.g. image segmentation and image parsing, has predominantly focused on the low-level visual clues within single image and neglected the contextual information across different images. In this paper, we present a new perspective to image decomposition piloted by the multi-labels associated with individual images. Observing that the context information (i.e., local label representations of the same label are similar while those from different labels are dissimilar) exists across different images, we propose to perform image decomposition in a collective way, and then the image decomposition problem is formulated as an optimization which maximizes inter-label difference and at the same time minimizes intra-label difference of the target label representations. Such contextual image decomposition has a wide variety of applications, among which the two exemplary ones are: 1) multi-label image annotation in which the sparse coding of a query image over the bases consisting of all learned label representations naturally produces the multi-label annotation, and 2) label ranking in which the annotated labels are re-ordered according to the sparse coding coefficients on those learned label representations. It is worth noting that these two applications can be performed simultaneously via the label propagation process in sparse coding.

On a large group picture, having all people open their eyes can turn out to be a difficult task for photographers. Therefore, in this paper, we describe an original method to automatically correct closed eyes on everyday pictures. For this aim, we explore the combination possibilities of (1) active shape model (ASM) to detect facial features, such as eyes, nose and head shape, and (2) Poisson editing to clone open eyes seamlessly. To improve the performance of seamless cloning, we suggest a pre-processing method that adjusts skin luminosity between two pictures. A nearest neighbor-based search to find the best suited pair of eyes among a set of donor candidates is also presented. We applied the proposed algorithm on several pictures and obtained very natural results, which demonstrates the validity of our approach.

A probabilistic matching of lines, which form a homography in two images, is formulated in the framework of the forward stepwise regression. A membership matrix represents the likelihood of line correspondences to the homography. The correspondence measure is borrowed from the forward stepwise regression so that the squared error of the homography and the number of correspondences are balanced simultaneously. An alternating scheme for optimizing the membership and homography is provided. The experimental results on synthetic and real images validate the proposed method.

Representing image using the distribution of local features on a group of visual words is an effective method for visual categorization. Visual words can be related conceptually and the information can be incorporated to enhance the performance. However, conventional methods usually use visual words independently without considering this. This paper proposes a novel approach to measure the conceptual relation of visual words and incorporate the information into visual categorization. The conceptual relation is measured by the similarity of class distributions induced by visual words, accordingly visual words are grouped and images are represented on multiple levels. Categorization is taken using the support vector machine (SVM) with an effective kernel designed for matching multi-level representations. The proposed method is evaluated for video events categorization on the benchmark dataset and shows superior performance to conventional methods.

In this paper, we present a local-driven semi-supervised learning framework to propagate the labels of the training data (with multi-label) to the unlabeled data. Instead of using each datum as a vertex of graph, we encode each extracted local feature descriptor as a vertex, and then the labels for each vertex from the training data are derived based on the context among different training data, finally the decomposed labels on each vertex are further propagated to the unlabeled vertices based on the similarities measured according to the features extracted at each local regions. With the learnt local descriptor graph we can predict the semantic labels for not only the test local features but also the test images. The experiments on multi-label image annotation demonstrate the encouraging results from our proposed framework of semi-supervised learning.

Unmanned Aerial Vehicles (UAV) are the subject of an increasing interest in many applications and a key requirement for autonomous navigation is the attitude/position stabilization of the vehicle. Some previous works have suggested using catadioptric vision, instead of traditional perspective cameras, in order to gather much more information from the environment and therefore improve the robustness of the UAV attitude/position estimation. This paper belongs to a series of recent publications of our research group concerning catadioptric vision for UAVs. Currently, we focus on the extraction of skyline in catadioptric images since it provides important information about the attitude/position of the UAV. For example, the DEM-based methods can match the extracted skyline with a Digital Elevation Map (DEM) by process of registration, which permits to estimate the attitude and the position of the camera. Like any standard cameras, catadioptric systems cannot work in low luminosity situations because they are based on visible light. To overcome this important limitation, in this paper, we propose using a catadioptric infrared camera and extending one of our methods of skyline detection towards catadioptric infrared images. The task of extracting the best skyline in images is usually converted in an energy minimization problem that can be solved by dynamic programming. The major contribution of this paper is the extension of dynamic programming for catadioptric images using an adapted neighborhood and an appropriate scanning direction. Finally, we present some experimental results to demonstrate the validity of our approach.

Shape is one of the useful information for object detection. The human visual system can often recognize objects based on the 2-D outline shape alone. In this paper, we address the challenging problem of shape matching in the presence of complex background clutter and occlusion. To this end, we propose a graph-based approach for shape matching. Unlike prior methods which measure the shape similarity without considering the relation among edge pixels, our approach uses the connectivity of edge pixels by generating a graph. A group of connected edge pixels, which is represented by an “edge” of the graph, is considered together and their similarity cost is defined for the “edge” weight by explicit comparison with the corresponding template part. This approach provides the key advantage of reducing ambiguity even in the presence of background clutter and occlusion. The optimization is performed by means of a graph-based dynamic algorithm. The robustness of our method is demonstrated for several examples including long video sequences. Finally, we applied our algorithm to our grasping robot system by providing the object information in the form of prompt hand-drawn templates.

A global pose estimation method of an Unmanned Aerial Vehicle (UAV) by matching forward-looking aerial images from the UAV flying at low altitude with down-looking images from a satellite is proposed. To overcome the limitation of significantly different camera viewpoints and characteristics, we use buildings as a cue of matching. We extract buildings from aerial images and construct a 3D model of buildings, using the fundamental matrix. We estimate the global pose of the vehicle by matching 3D structure of buildings with satellite images, using a particle filter. Experimental results show that the proposed approach is a promising method to the global pose estimation of the UAV with forward-looking vision data.

Region descriptor has proved to be very important for local feature based image categorization. Previous region descriptors are usually based on the statistics of low level features, such as intensity, edge response, and etc. In this paper a novel descriptor named local texton statistics (LTS) that explores the high level semantic statistical characteristics of image regions is presented. Perceptual information is obtained by applying Gaussian filter banks and the image regions are described by the statistics of different 'texton's. Using the bag of words as classification algorithm, experiments show that the proposed descriptor is superior to the previous popular SIFT descriptors on the Wang dataset. The combination of these two descriptors shows high performance for categorization on both the Wang dataset and the fifteen scene categories dataset.

In this paper, feature point matching is formulated as an optimization problem in which the uniqueness condition is constrained. We propose a novel score function based on homography-induced pairwise constraints, and a novel optimization algorithm based on relaxation labeling. Homography-induced pairwise constraints are effective for image pairs with viewpoint or scale changes, unlike previous pairwise constraints. The proposed optimization algorithm searches for a uniqueness-constrained solution, while the original relaxation-labeling algorithm is appropriate for finding many-to-one correspondences. The effectiveness of the proposed method is shown by experiments involving image pairs with viewpoint or scale changes in addition to repeated textures and nonrigid deformation. The proposed method is also applied to object recognition, giving some promising results.

Most of vision-based algorithms for motion and localization estimation requires matching some interest points in a pair of images. After building feature correspondence, it is possible to estimate camera motion/localization using epipolar geometry. However feature matching is still a challenging problem because of time constraint or image variability for example. In several robotic applications, the camera rotation may be known thanks to a gyroscope or another orientation sensor. Therefore, in this paper, we aim to answer the following question: can the knowledge of rotation from a gyroscope be used to improve feature matching. To analyze this new approach of camera and gyroscope data fusion, we proceed in two steps. First, we rotationally align the images using rotation information of the gyroscope. And second, we compare the quality of feature matching in the original and rotationally aligned images. Experimental results on a real catadioptric sequence show that gyroscope data permits to sensibly improve the number of inliers according to epipolar geometry.

In structure from motion, if two metric structures are given, the unknown scale between them can be resolved by constraining the rigidity of the metric space. There exist two well-known approaches. The first one is the pose estimation which aims to find the pose of a camera for known 3D points. The second one is the scale estimation whose goal is to resolve the scale after estimating the motion between cameras. Recently, the former way is preferred because of the vulnerability of the scale estimation, the weakness to the image noise. In this paper, we thus propose a robust method which can overcome the weakness of the scale estimation by considering the uncertainty of reconstructed 3D points. Additionally, the rotation matrix is directly corrected under the structural consistency constraint. To illustrate the performance of the method, we demonstrate some examples of large-scale reconstructions and compare the results.

Finding correspondences of two images taken from largely different camera configuration is a challenging problem because appearance information such as color, intensity and edge orientation histogram cannot be used. A probabilistic approach to match line segments in the images is proposed for calculating their homography. A membership matrix to represent contribution of every match to the homography is employed. A relaxed version of forward stepwise regression is derived and the test statistic is optimized. An alternating scheme for optimizing the membership and homography is provided. The simulation results on synthetic images validate the proposed method.

Unmanned aerial vehicles (UAV) are the subject of an increasing interest in many applications and a key requirement is the stabilization of the vehicle. Some previous works have suggested using catadioptric vision, instead of traditional perspective cameras, in order to gather much more information from the environment and therefore improve the robustness of the UAV attitude estimation. This paper belongs to a series of recent publications of our research group concerning catadioptric vision for UAVs. Currently, we focus on the estimation of the complete attitude of a UAV flying in urban environment. In order to avoid the limitations of horizon-based approaches, the difficulties of traditional epipolar methods (such as rotation-translation ambiguity, lack of features, retrieving motion parameters from matrix decomposition, etc..) and improve UAV dynamic control, we suggest computing infinite homography. We show how catadioptric vision plays a key role to: first, extract a large number of lines, second robustly estimate the associated vanishing points and third, track them even during long video sequences. Therefore it is not only possible to estimate the relative rotation between consecutive frames but also compute the absolute rotation between two distant frames without error accumulation. Finally, we present some experimental results with ground truth data to demonstrate the accuracy and the robustness of our method.

A key requirement for unmanned aerial vehicles (UAV) applications is the attitude stabilization of the aircraft, which requires the knowledge of its orientation. It is now well established that traditional navigation equipments, like GPS or INS, suffer from several disadvantages. That is why some works have suggested a vision-based approach of the problem. Especially, catadioptric vision is more and more used since it permits to gather much more information from the environment, compared to traditional perspective cameras, and therefore the robustness of the UAV attitude estimation is improved. Rotation estimation from conventional and catadioptric images has been extensively studied. Whereas interesting results can be obtained, the existing methods have non-negligible limitations such as difficult features matching (e.g. repeated texture, blurring or illumination changing) or a high computational cost (e.g. vanishing point extraction or analyze in frequency domain). In order to overcome these limitations, this paper presents a top-down approach for estimating the rotation and extracting the vanishing points in catadioptric images. This new framework is accurate and can run in real-time. To obtain the ground truth data, we also calibrate our catadioptric camera with a gyroscope. Finally, experimental results on a real video sequence are presented and compared to the ground truth data obtained by the gyroscope.

Camera calibration is an important step for vision-based stabilization of unmanned aerial vehicles (UAV). The goal of this paper is to develop a method for automatic calibration of a catadioptric camera so that it can be easily run before mounting the camera on the UAV or even during the flight to deal with vibrations or shocks. Whereas existing works can provide interesting results, they suffer from several practical limitations (manual line extraction, inaccurate conic fitting, calibration pattern, camera motion, execution time, etc...) and therefore cannot be applied in our application. The proposed algorithm aims to determine the most probable calibration that verifies some geometric constraints induced by catadioptric projection. In order to efficiently maximize this probability, we use a particle filtering approach. Experimental results have demonstrated the effectiveness of the proposed method.

Feature extraction and matching is one of the most significant research areas in robot vision. In this paper, we present a new method for motion estimation and mapping using color feature extraction and matching. The proposed method reduces computational cost and has good performance. The experimental result shows that the proposed method not only runs faster but provides accurate result.

Recently, many vision-based navigation methods have been introduced as an intelligent robot application. However, many of these methods mainly focus on finding an image in the database corresponding to a query image. Thus, if the environment changes, for example, objects moving in the environment, a robot is unlikely to find consistent corresponding points with one of the database images. To handle these problems, we propose a novel motion-based navigation method in contrast with appearance-based approaches. This algorithm is based on motion estimation by a camera to plan the next movement of a robot and robust feature matching to recognize home and destination locations. Experimental results demonstrate the capability of the vision-based autonomous navigation against environment changes.

Recently, many vision-based robotic applications such as visual SLAM (Simultaneous Localization And Mapping) and autonomous navigation have achieved good performance using visual features. In these applications, robust feature tracking plays an important role, e.g., in scene recognition for autonomous navigation and in data association for visual SLAM. In this paper, we propose a hierarchical outlier detection algorithm for robust feature tracking; the algorithm uses a simple window-based correlation (NCC) and enforces angular and scale constraints. The proposed algorithm maximizes the inter-cluster score and detects outliers that do not satisfy the angular constraints. The remaining outliers are detected by enforcing scale constraints using SIFT descriptors. The proposed algorithm is efficient and particularly useful for scene recognition, in which an image corresponding to a query image is searched among data images. Experimental results demonstrate that the proposed algorithm is robust to outliers and image variations such as scale changes. One of the main applications of the proposed algorithm is global localization due to its low computational complexity and robustness to outliers.

Representation of images using the distribution of local features on a visual codebook is an effective method for object categorization. Typically, discriminative capability of the codebook can lead to a better performance. However, conventional methods usually use clustering algorithms to learn codebooks without considering this. This paper presents a novel approach of learning optimal compact codebooks by selecting a subset of discriminative codes from a large codebook. Firstly, the Gaussian models of object categories based on a single code are learned from the distribution of local features within each image. Then two discriminative criteria, i.e. likelihood ratio and Fisher, are introduced to evaluate how each code contributes to the categorization. We evaluate the optimal codebooks constructed by these two criteria on Caltech-4 dataset, and report superior performance of object categorization compared with traditional K-means method with the same size of codebook.

In this paper, we introduce the Skellam distribution as a sensor noise model for CCD or CMOS cameras. This is derived from the Poisson distribution of photons that determine the sensor response. We show that the Skellam distribution can be used to measure the intensity difference of pixels in the spatial domain, as well as in the temporal domain. In addition, we show that Skellam parameters are linearly related to the intensity of the pixels. This property means that the brighter pixels tolerate greater variation of intensity than the darker pixels. This enables us to decide automatically whether two pixels have different colors. We apply this modeling to detect the edges in color images. The resulting algorithm requires only a confidence interval for a hypothesis test, because it uses the distribution of image noise directly. More importantly, we demonstrate that without conventional Gaussian smoothing the noise model-based approach can automatically extract the fine details of image structures, such as edges and corners, independent of camera setting.

Nowadays, robotic systems are more and more equipped with catadioptric cameras. However several problems associated to catadioptric vision have been studied only slightly. Especially algorithms for detecting rectangles in catadioptric images have not yet been developed whereas it is required in diverse applications such as building extraction in aerial images. We show that working in the equivalent sphere provides an appropriate framework to detect lines, parallelism, orthogonality and therefore rectangles. Finally, we present experimental results on synthesized and real data.

The point ambiguity owing to the ambiguous local appearances of image points is the one of the main causes making the stereo problem difficult. Under the point ambiguity, local similarity measures are easy to be ambiguous and this results in false matches in ambiguous regions. In this paper, we present the new similarity measure to resolve the point ambiguity problem based on the idea that the distinctiveness, not the interest, is the appropriate criterion for the feature selection under the point ambiguity. The proposed similarity measure named the Distinctive Similarity Measure (DSM) is essentially based on the distinctiveness of image points and the dissimilarity between them, which are both closely related to the local appearances of image points; the distinctiveness of an image point is related to the probability of a mismatch while the dissimilarity is related to the probability of a good match. We verify the efficiency of the proposed DSM by using testbed image sets. Experimental results show that the proposed DSM is very effective and can be easily used for improving the performance of existing stereo methods under the point ambiguity.

The CCD camera and the 2D laser range finder are widely used for motion estimation and 3D reconstruction. With their own strengths and weaknesses, low-level fusion of these two sensors complements each other. We combine these two sensors to perform motion estimation and 3D reconstruction simultaneously and precisely. We develop a motion estimation scheme appropriate for this sensor system. In the proposed method, the motion between two frames is estimated using three points among the scan data, and refined by nonlinear optimization. We validate the accuracy of the proposed method using real images. The results show that the proposed system is a practical solution for motion estimation as well as for 3D reconstruction.

Categorizing visual elements is fundamentally important for autonomous mobile robots to get intelligence such as new object acquisition and topological place classification. The main problem of visual categorization is how to reduce the large intra-class variations, especially surface markings of man-made objects. In this paper, we present a robust method by introducing intermediate blurring and entropy-guided codebook selection in a bag-of-words framework. Intermediate blurring can filter out the high frequency of surface markings and provide dominant shape information. Entropy of a hypothesized codebook can provide the necessary measure for the semantic parts among training exemplars. From the first step, a generative optimal codebook for each category is learned using the MDL (minimum description length) principle guided by entropy information. From the second step, a final set of codebook is learned using the discriminative method guided by the inter-category entropy of the codebook. We select the necessary parameters through various evaluations and validate the effect of the surface marking reduction method using a Caltech-101 DB, which has large intra-class variations. Finally, we briefly introduce the impact of the method to the object categorization and segmentation problem

Recently, many vision-based SLAM methods have achieved good results using visual features. However, most algorithms suffer from the accumulated error that inevitably occurs. In this paper, we propose a robust loop detection method by matching image features between the incoming image and key-frame images saved in SLAM. Loop detection is a task of deciding whether a robot has returned to a previously visited area or not. Because a camera is unlikely to have the same pose when a robot revisits the place where it previously encountered, it is crucial to match the features under the different views of the scene. In contrast with view-invariant features, it is hard to match corner points in that situation due to the large variation of neighboring pixels. So we present the robust corner matching method under the view changes. Experimental results demonstrate the capability of the loop closing and mobile robot localization under the different views using the proposed method.

Visual categorization is fundamentally important for autonomous mobile robots to get intelligence such as novel object acquisition and topological place recognition. The main difficulty of visual categorization is how to reduce the large intra-class variations. In this paper, we present a new method made robust to that problem by using intermediate blurring and entropy-guided codebook selection in a bag-of-words framework. Intermediate blurring can reduce the high frequency of surface markings and provide dominant shape information. Entropy of a hypothesized codebook can provide the necessary amount of repetition among training exemplars. A generative optimal codebook for each category is learned using the MDL (minimum description length) principle guided by entropy information. Finally, a discriminative codebook is learned using the discriminative method guided by the inter-category entropy of the codebook. We validate the effect of the proposed method using a Caltech-101 DB, which has large intra-class variations

In this paper, we present a practical place and object recognition method for guiding visitors in building environments. Due to motion blur or camera noise, places or objects can be ambiguous. The first key contribution of this work is the modeling of bidirectional interaction between places and objects for simultaneous reinforcement. The second key contribution is the unification of visual context, including scene context, object context, and temporal context. The last key contribution is a practical demonstration of the proposed system for visitors in a large scale building environment. 1

Plane-based calibration is now a very popular procedure because of its flexibility. One key step consists in detecting a set of coplanar features, from which the Euclidean structure of the corresponding 3D plane has to be computed. We suggest to use confocal conics as calibration targets, as they offer undeniable advantages over other ones (e.g., points or lines) in terms of detection and estimation, especially in the presence of partial occlusion. We introduce important projective and Euclidean properties of the linear family of conics (i.e., the confocal conic range), spanned by two confocal conics. In particular, we rely on the fact that the circular point-envelope - a rank-2 conic that encodes the 2D Euclidean structure - is a degenerate member of any confocal conic range. This allows us to give closed-form solutions in three cases: one conic with known foci, two confocal conics with known product of ratios of semi axes, and two unknown confocal conics. The performances of the proposed algorithms (consisting of a few lines of Matlab-like code) show up high accuracies for both intrinsic and extrinsic camera parameters. In addition to experiments with synthetic data, a video sequence is processed, showing off the interest of using confocal conics as calibration targets, for augmented reality purposes.

Recently, many global stereo methods have achieved good results by modeling a disparity surface as a Markov random field (MRF) and by solving an optimization problem with various techniques. However, most global methods mainly focus on how to minimize conventional cost functions efficiently, although it is more important to define cost functions well to improve performance. In this paper, we propose new symmetric cost functions for global stereo methods. We first present a symmetric data cost function for the likelihood and then propose a symmetric discontinuity cost function for the prior in the MRF model for stereo. In defining cost function, both the reference image and the target image are taken into account to improve performance without modeling half-occluded pixels explicitly and without using color segmentation. The performance improvement of stereo matching due to the proposed symmetric cost functions is verified by applying the proposed symmetric cost functions to the belief propagation (BP) based stereo method. Experimental results for standard testbed images show that the performance of the BP based stereo method is greatly improved by the proposed symmetric cost functions.

Most change detection methods are based on gray-level images. A gray-level image is regarded as a 1-D projection of three channels of color images. Therefore, more precise change detection results are expected by utilizing color information. We previously developed a change detection scheme using color images. In this paper, we determine which color space should be selected for accurate change detection based on our previous detection scheme. Our method can be applied to various color spaces, including gray-level images. Then we can measure the expected number of error pixels in order to select an appropriate color space which gives the best result among various color spaces. The experiments show that selecting a color space based on measurements results in the fewest error pixels.

In this paper, we propose a fast method for separating reflection components using a single color image. We first propose a specular-free two-band image that is a specularity-invariant color image representation. Reflection components separation is achieved by comparing local ratios at each pixel and making those ratios equal in an iterative framework. The proposed method is very fast and shows reasonable results for textured indoor/outdoor images.

We present a new window-based method for correspondence search using varying support-weights. We adjust the support-weights of the pixels in a given support window based on color similarity and geometric proximity to reduce the image ambiguity. Our method outperforms other local methods on standard stereo benchmarks.

In this paper, we present a new area-based method for visual correspondence search that focuses on the dissimilarity computation. Local and area-based matching methods generally measure the similarity (or dissimilarity) between the image pixels using local support windows. In this approach, an appropriate support window should be selected adaptively for each pixel to make the measure reliable and certain. Finding the optimal support window with an arbitrary shape and size is, however, very difficult and generally known as an NP-hard problem. For this reason, unlike the existing methods that try to find an optimal support window, we adjusted the support-weight of each pixel in a given support window. The adaptive support-weight of a pixel is computed based on the photometric and geometric relationship with the pixel under consideration. Dissimilarity is then computed using the raw matching costs and support-weights of both support windows, and the correspondence is finally selected by the WTA (winner-takes-all) method. The experimental results for the rectified real images show that the proposed method successfully produces piecewise smooth disparity maps while preserving sharp depth discontinuities accurately.

In this paper, we present a new dichromatic-based method for estimating illuminant chromaticity using a single image. In the proposed method, specular pixels are roughly extracted from an input image and projected into the three-dimensional dichromatic line space according to their image chromaticities and dichromatic slopes. Reliable dichromatic lines are then obtained by detecting local maxima in the dichromatic line space. Illuminant chromaticity is finally estimated by finding the intersections of the resulting dichromatic lines. The proposed method has some advantages. The proposed method can deal with highly textured images since it does not require precise color segmentation. In addition, the proposed method does not suffer from nearly parallel dichromatic lines and it can estimate artificial illuminant chromaticity.

The image sequence of a static scene includes similar or redundant information over time. Hence, motion-discontinuous instants can efficiently characterize a video shot or event. However, such instants (key frames) are differently identified according to the change of velocity and acceleration of motion, and such scales of change might be different on each sequence of the same event. In this paper, we present a scalable video abstraction in which the key frames are obtained by the maximum curvature of camera motion at each temporal scale. The scalability means dealing with the velocity and acceleration change of motion. In the temporal neighborhood determined by the scale, the scene features (motion, color, and edge) can be used to index and classify the video events. Therefore, those key frames provide temporal interest points (TIPs) for the abstraction and classification of video events.

We present an accurate metric localization method using a simple artificial landmark for the navigation of indoor mobile robots. The proposed landmark model is designed to have a three-dimensional, multi-colored structure and the projective distortion of the structure encodes the distance and heading of the robot with respect to the landmark. Catadioptric vision is adopted for the robust and easier acquisition of the bearing measurements for the landmark. We propose a practical EKF based self-localization method that uses a single artificial landmark and runs in real time.

In this paper, we present a recognition-based autonomous navigation system for mobile robots. The system is based on our previously proposed robust invariant feature (RIF) detector. This detector extracts highly robust and repeatable features based on the key idea of tracking multi-scale interest points and selecting unique representative local structures with the strongest response in both spatial and scale domains. Weighted Zernike moments are used as the feature descriptor and applied to the place recognition. The navigation system is composed of on-line and off-line two stages. In the off-line learning stage, we train the robot in its workspace by just taking several images of representative places as landmarks. Then, in the on-line navigation stage, the robot recognizes scenes, obtains robust feature correspondences, and navigates the environment autonomously using the iterative pose converging (IPC) algorithm which is based on the idea of the visual servoing technique. The experimental results and the performance evaluation show that the proposed navigation system can achieve excellent performance in complex indoor environments.

This paper presents a combined model-based 3D object recognition method motivated by the robust properties of human vision. The human visual system (HVS) is very efficient and robust in identifying and grabbing objects, in part because of its properties of visual attention, contrast mechanism, feature binding, multiresolution and part-based representation. In addition, the HVS combines bottom-up and top-down information effectively using combined model representation. We propose a method for integrating these aspects under a Monte Carlo method. In this scheme, object recognition is regarded as a parameter optimization problem. The bottom-up process initializes parameters, and the top-down process optimizes them. Experimental results show that the proposed recognition model is feasible for 3D object identification and pose estimation.

We investigate the projective properties of the feature consisting of two concentric circles. We demonstrate there exist geometric and algebraic constraints on its projection. We show how these constraints greatly simplify the recoveries of the affine and Euclidean structures of a 3D plane. As an application, we assess the performances of two camera calibration algorithms.

We present a novel change detection method using a statistical model of the image noise. Most change detection methods are based on gray-level images. However, color images can provide much richer scene information. One major problem to use the color images in change detection is how to combine three components in color space as a detection cue. We use the Euclidean color distance of three channels to measure the difference between two consecutive images. Specifically, we present a new noise model for each color channel. Through this modeling we can estimate the distribution of the Euclidean color distance for unchanged regions. We can find the optimal threshold to detect changes using this estimated distribution. Although we use the optimal threshold, inevitably there may be false classifications. To reject these erroneous cases, we adopt the graph cuts method that efficiently minimizes the global energy, which takes into account the effect of neighboring pixels.

We propose a vision based 3D object recognition and tracking system, which provides high level scene descriptions such as object identification and 3D pose information. The system is composed of object recognition part and real-time tracking part. In object recognition, we propose a feature which is robust to scale, rotation, illumination change and background clutter. A probabilistic voting scheme maximizes the conditional probability defined by the features in correspondence to recognize an object of interest. As a result of object recognition, we obtain the homography between the model image and the input scene. In tracking, a Lie group formalism is used to cast the motion computation problem into simple geometric terms so that tracking becomes a simple optimization problem. An initial object pose is estimated using correspondences between the model image and the 3D CAD model which are predefined and the homography which relates the model image to the input scene. Results from the experiments show the robustness of the proposed system.

This paper proposes a new edge detection method using a 3/spl times/3 ideal binary pattern and lookup table (LUT) for the mobile robot localization without any parameter adjustments. We take the mean of the pixels within the 3/spl times/3 block as a threshold by which the pixels are divided into two groups. The edge magnitude and orientation are calculated by taking the difference of average intensities of the two groups and by searching directional code in the LUT, respectively. And also the input image is not only partitioned into multiple groups according to their intensity similarities by the histogram, but also the threshold of each group is determined by fuzzy reasoning automatically. Finally, the edges are determined through non-maximum suppression using edge confidence measure and edge linking. Applying this edge detection method to the mobile robot localization using projective invariance of the cross ratio, we demonstrate the robustness of the proposed method to the illumination changes in a corridor environment.

We present a simple artificial landmark model and a robust tracking algorithm for the navigation of indoor mobile robots. The landmark model is designed to have a three-dimensional structure consisting of a multi-colored planar pattern. A stochastic algorithm based on Condensation [1] tracks the landmark model robustly using the color distribution of the pattern. A new self-localization algorithm computes the location of robot with the tracked single landmark. Experimental results show that the proposed landmark model is eflective. Through extensive navigation experiments in a cluttered indoor environment, we demonstrate the feasibility of the single view based self-localization in real-time.

In this paper we present a new robust face tracking method based on the condensation algorithm that uses a sampling based density representation. A two-dimensional color model is used to approximate the face color. We modified the condensation algorithm to provide color adaptability to the abrupt change of illumination and to the tracking of differently colored people. According to the face size and location uncertainty, the searching range is automatically determined and it makes the algorithm extremely robust and efficient. The tracker operates at real-time and actively controls a camera pan-tilt in order to locate a person's face in the center of the image. Experimental results show the algorithm's robustness to the agile motion of face and to the dramatic change of illumination in the presence of complex background.

In this paper, we present a camera calibration method using two views of three pairs of concentric circles with known sizes. A new invariant property for circle is introduced to determine the position of the center of the projected circle. Given two image ellipses and their corresponding centers, a cross-ratio based method estimates the center of projected circles using the new invariant property. The accurate center of projected concentric circles provides correct correspondences between ellipses in the image plane and circles in 3D plane. We also demonstrate that two views of three pairs of coplanar concentric circles are enough to determine the intrinsic camera parameters, such as the focal length, the aspect ratio, and the principal point. We validate the performance of the method using both synthetic and real images. Our method shows a comparable performance with respect to similar calibration methods using planes. The use of concentric circles, however, provides correct correspondences between 3D target points and their image points, and greatly simplifies the calibration problem.

For the fast and accurate self-localization of mobile robots, landmarks can be used very efficiently in the complex workspace. In this paper, we propose a simple color landmark model for self-localization and a fast landmark detection and tracking algorithm based on the proposed landmark model. We develop a color landmark with symmetric and repetitive structures, which shows invariant color histogram characteristics under some geometric distortions. Detection and tracking of the model are accomplished by a factored sampling technique in which color similarity is estimated by the color histogram intersection. We also use the color similarity to update the color histogram model of the landmark model for robust tracking under illumination change. We demonstrate the feasibility of the proposed technique through experiments in cluttered indoor environments.

Self-calibration algorithms that use only the information in the image have been actively researched. However, most algorithms require bundle adjustment in the projective reconstruction or in the nonlinear minimization. We propose a practical self-calibration algorithm that only requires a linear projective reconstruction. We overcome the sensitivity of the algorithm due to image noises by adding another constraint on the principal point. Also, we propose a variant of linear auto-calibration algorithm which uses the similar assumption of the work of Pollefeys et al. (1998), based on the property of the absolute quadric. Experimental results using real and synthetic images demonstrate the feasibility of the proposed algorithm.

In this paper, we analyze and solve the problem of image-based visual servoing system under large initial pose error. In this case we must consider camera field of view and the violation of the assumption of image-based visual servoing. Control-law is derived from linear approximation of the image error between the gripper and the target object. To solve this problem, we propose a unified image-based control strategy with a new image space named as virtual image plane (VIP). And next we propose an image-based control-law by decoupling the translation and rotation motion of the large pose error The remarkable property of our method is that we only use the combination of the image features in the 2D image space. Finally we demonstrate the feasibility of the proposed system through experiments using a six-DOF robot equipped with a stereo camera.

We propose a novel and practical stereo camera system that uses only one camera and a biprism placed in front of the camera. The equivalent of a stereo pair of images is formed as the left and right halves of a single charge coupled device (CCD) image using a biprism. The system is therefore cheap and extremely easy to calibrate since it requires only one CCD camera. An additional advantage of the geometrical setup is that corresponding features lie on the same scanline automatically. The single camera and biprism have led to a simple stereo system for which correspondence is very easy and accurate for nearby objects in a small field of view. Since we use only a single lens, calibration of the system is greatly simplified. Given the parameters in the biprism-stereo camera system, we can reconstruct the three-dimensional structure using only the disparity between the corresponding points.

In this paper we propose a novel and practical stereo camera system that uses only one camera and a biprism placed in front of the camera. The equivalent of a stereo pair of images is formed as the left and right halves of a single CCD image using a biprism. The system is therefore cheap and extremely easy to calibrate since it requires only one CCD camera. An additional advantage of the geometrical set-up is that corresponding features lie on the same scanline automatically. The single camera and biprism have led to a simple stereo system for which correspondence is very easy and which is accurate for nearby objects in a small field of view. Since we we only, a single lens, calibration of the system is greatly simplified. This is due to the fact that we need to estimate only one focal length and one center of projection. Given the parameters in the biprism-stereo camera system, we can recover the depth of the object using only the disparity between the corresponding points.

We present an algorithm for the calibration of a camera and the recovery of 3D scene structure up to a scale from image sequences using known angles between lines in the scene. The proposed method computes the intrinsic parameters of the camera using the invariance of angles under the similarity transformation. Specifically, we recover the matrix that is the homography between the projective structure and the Euclidean structure using angles. Since this matrix is a unique one in the given set of image sequences, we can easily deal with the problem of varying intrinsic parameters of the camera. Experimental results on the synthetic and real images demonstrate the feasibility of the proposed algorithm.

We propose a new method for recognizing three-dimensional objects using a three-dimensional invariant relationship for a special structure and geometric hashing by single-view. We use a special structure consisting of four co-planar points and any two non-coplanar points with respect to the plane. We derive an invariant relationship for the structure, which is represented by a plane equation. For recognition of 3-D objects using geometric hashing, a set of points on the plane is mapped into a set of points intersecting the plane and the unit sphere, thereby satisfying the invariant relationship. Experiments using 3-D polyhedral objects are carried out to demonstrate the feasibility of our method for 3-D object recognition.

This paper presents an effective recognition method based on perceptual organization of low level features detected in an image. The method uses a dynamic programming (DP) based formulation to represent various line groups such as convex, concave, and more complex patterns consisting of convex and concave shapes. The essential features of perceptual organization such as endpoint proximity, collinearity, parallelism, and connectivity of lines, are incorporated into the DP based formulation as energy terms. As endpoint proximity, we detect two line junctions from image lines. We then search for junction groups by using collinearity constraint between the junctions. A DP-based search algorithm is used to detect a junction chain similar to the model chain, based on a local comparison. The proposed system is able to find line groups from images with broken lines and strong background clutters. We demonstrate the feasibility of our DP-based matching method based on perceptual organization using real images.

This paper proposes a segmentation algorithm by means of an evidential reasoning to segment moving vehicles in front of our moving car in a road traffic scene. Generally, an evidential reasoning finds the perceptually known evidences of a target and updates a probabilistic expectation for the target to be in an image. Since a noise image produces unreliable features and degrades the detection and localization, selecting image primitives which are less sensitive to noise and well represent the evidences is important. We carry out this task by the probabilistic integration of image features based on maximum a posteriori (MAP) probability that combines the prior and likelihood probabilities using Bayes' rule.

In this paper, we propose two new vision-based methods for indoor mobile robot navigation. One is a self-localization algorithm using projective invariant and the other is a method for obstacle detection by simple image difference and relative positioning. For a geometric model of corridor environment, we use natural features formed by floor, walls, and door frames. Using the cross-ratios of the features can be effective and robust in building and updating model-base, and image matching. We predefine a risk zone without obstacles for a robot, and store the image of the risk zone, which will be used to detect obstacles inside the zone by comparing the stored image with the current image of a new risk zone. The position of the robot and obstacles are determined by relative positioning. The robustness and feasibility of our algorithms have been demonstrated through experiments in corridor environments using the KASIRI-II indoor mobile robot.

This paper presents a robust vision algorithm for tracking the boundary of an object with an arbitrary shape by using monocular image sequences. This method consists of a curve registration based optimization technique and a deformable contour model ("snakes") for the global and the local motion estimations, respectively. By combining techniques, we overcome, among other problems, inaccurate estimate of motion parameters in the curve registration method (which apparently only occur when a rigid or a flexible object is tracked), and the "local position variation" of the deformable contour model, variations, which are due to noisy images and/or complex backgrounds. The curve registration method uses an iterative algorithm to find the minimum normal distance between two curves, one before motion and the corresponding curve after it. Snakes overcome the limitation of the curve registration method, which suffers from the inaccuracy of motion models. We also propose an internal force, which increases local robustness of the deformable contour to background noise. By using the refined snakes' control points, the global update of the previous curve is performed for the re-location of the registered curve. Additionally, we integrate the geometric invariant value of the boundary contour and the curve registration method to solve the occlusion problem in visual tracking. The proposed method is validated through experiments on real images.

Robust and effective real-time visual tracking is realized by combining the first order differential invariants with stochastic filtering. The Kalman filter as an optimal stochastic filter is used to estimate the motion parameters, namely the plant state vector of the moving object with the unknown dynamics in successive image frames. Using the fact that the relative motion between the moving object and the moving observer causes the deformation, we compute the first differential invariants of the image velocity field. The surface orientation and the depth estimate between the observer and the object are computed based on these first order differential invariants. We demonstrate the robustness and feasibility of the proposed tracking algorithm through real experiments in which an X-Y Cartesian robot tracks a toy vehicle moving along 3D rails.

The authors present a navigation system using multiple sensors for unknown and dynamic indoor environments. To achieve robustness and flexibility in the mobile robot, a behavior-based system architecture is developed consisting of multilayered behaviors using multiple sensors that were ultrasonic sensors and a video camera. Basic behaviors required for navigation, such as avoiding obstacles, moving toward free space, and following targets are redundantly developed as agents and combined in a behavior-based system architecture. The capabilities of the system were demonstrated in unstructured real office environments using an indoor mobile robot.>

The authors are developing a compact cart-type robot system which moves around an office. They propose a behavior-based architecture with three clustered (reflexive, purposive, and adaptive) agents that realizes efficiency in attaining the mission of the robot, and robustness against the various kinds of failures that may occur in a dynamic environment. The reflexive-level group consists of agents with contact, infrared, and ultrasonic sensors which maintain minimal safety of the robot. The role of the purposive-level group is to achieve the global mission of the robot, such as 'if the robot detects a small fire in the office, find and reach a fire extinguisher as soon as possible'. The adaptive-level group stands by to recover from failure in the purposive-level group or in a deadlock situation. Experimental results showed the effectiveness of the method.>

We present a navigation system using multiple sensors for unknown and dynamic indoor environments. To achieve the robustness and flexibility of the mobile robot, we propose a new behavior-based architecture with three groups of clustered (reflexive, purposive, and adaptive) agents that realizes both efficiency in attaining the mission of the robot and robustness against the various kinds or failures that may occur in a dynamic environment. Basic behaviors required for navigation, such as, avoiding obstacles, moving towards free space, and following targets, are redundantly developed as agents and combined in the behavior-based system architecture.