ECCV 2020 论文列表

Retrieving content relevant images from a large-scale fine-grained dataset could suffer from intolerably slow query speed and highly redundant storage cost, due to high-dimensional real-valued embeddings which aim to distinguish subtle visual differences of fine-grained objects. In this paper, we study the novel fine-grained hashing topic to generate compact binary codes for fine-grained images, leveraging the search and storage efficiency of hash learning to alleviate the aforementioned problems. Specifically, we propose a unified end-to-end trainable network, termed as ExchNet. Based on attention mechanisms and proposed attention constraints, it can firstly obtain both local and global features to represent object parts and whole fine-grained objects, respectively. Furthermore, to ensure the discriminative ability and semantic meaning's consistency of these part-level features across images, we design a local feature alignment approach by performing a feature exchanging operation. Later, an alternative learning algorithm is employed to optimize the whole ExchNet and then generate the final binary hash codes. Validated by extensive experiments, our ExchNet consistently outperforms state-of-the-art generic hashing methods on five fine-grained datasets, which shows our effectiveness. Moreover, compared with other approximate nearest neighbor methods, ExchNet achieves the best speed-up and storage reduction, revealing its efficiency and practicality."

Example-guided image synthesis has recently been attempted to synthesize an image from a semantic label map and an exemplary image. In the task, the additional exemplar image provides the style guidance that controls the appearance of the synthesized output. Despite the controllability advantage, the existing models are designed on datasets with specific and roughly aligned objects. In this paper, we tackle a more challenging and general task, where the exemplar is a scene image that is semantically different from the given label map. To this end, we first propose a Masked Spatial-Channel Attention (MSCA) module which models the correspondence between two scenes via efficient decoupled attention. Next, we propose an end-to-end network for joint global and local feature alignment and synthesis. Finally, we propose a novel self-supervision task to enable training. Experiments on the large-scale and more diverse COCO-stuff dataset show significant improvements over the existing methods. Moreover, our approach provides interpretability and can be readily extended to other content manipulation tasks including style and spatial interpolation or extrapolation."

A domain adaptive object detector aims to adapt itself to unseen domains that may contain variations of object appearance, viewpoints or backgrounds. Most existing solutions adopt feature alignment either on the image level or instance level. However, image-level alignment on global features may tangle foreground/background pixels at the same time, while instance-level alignment using proposals may suffer from the background noise. Different from existing solutions, we propose a domain adaptation framework that accounts for each pixel, especially via predicting pixel-wise objectness and centerness. Specifically, the proposed method carries out center-aware alignment by paying more attention to foreground pixels, hence achieving better adaptation across domains. We demonstrate our method on numerous adaptation settings with extensive experimental results and show favorable performance against existing state-of-the-art algorithms."

Event cameras are bio-inspired sensors that respond to per-pixel brightness changes in the form of asynchronous and sparse “events”. Recently, pattern recognition algorithms, such as learning-based methods, have made significant progress with event cameras by converting events into synchronous dense, image-like representations and applying traditional machine learning methods developed for standard cameras. However, these approaches discard the spatial and temporal sparsity inherent in event data at the cost of higher computational complexity and latency. In this work, we present a general framework for converting models trained on synchronous image-like event representations into asynchronous models with identical output, thus directly leveraging the intrinsic asynchronous and sparse nature of the event data. We show both theoretically and experimentally that this drastically reduces the computational complexity and latency of high-capacity, synchronous neural networks without sacrificing accuracy. In addition, our framework has several desirable characteristics: (i) it exploits spatial and temporal sparsity in the event data explicitly, (ii) it is agnostic to the event representation, network architecture, and task, and (iii) it does not require any train-time change, since it is compatible with the standard neural net-works’ training process. We thoroughly validate the proposed framework on two computer vision tasks: object detection and object recognition. In these tasks, we reduce the computational complexity up to 20 times with respect to high-latency neural networks. At the same time, we outperform state-of-the-art asynchronous approaches up to 24.5% in prediction accuracy."

With extremely high temporal resolution, event cameras have a large potential for robotics and computer vision. However, their asynchronous imaging mechanism often aggravates the measurement sensitivity to noises and brings a physical burden to increase the image spatial resolution. To recover high-quality intensity images, one should address both denoising and super-resolution problems for event cameras. Since events depict brightness changes, with the enhanced degeneration model by the events, the clear and sharp high-resolution latent images can be recovered from the noisy, blurry and low-resolution intensity observations. Exploiting the framework of sparse learning, the events and the low-resolution intensity observations can be jointly considered. Based on this, we propose an explainable network, an event-enhanced sparse learning network (eSL-Net), to recover the high-quality images from event cameras. After training with a synthetic dataset, the proposed eSL-Net can largely improve the performance of the state-of-the-art by 7-12 dB. Furthermore, without additional training process, the proposed eSL-Net can be easily extended to generate continuous frames with frame-rate as high as the events. "

Modern methods for counting people in crowded scenes rely on deep networks to estimate people densities in individual images. As such, only very few take advantage of temporal consistency in video sequences, and those that do only impose weak smoothness constraints across consecutive frames.

Optimization-based image smoothing is routinely formulated as the game between a smoothing energy and an appearance preservation energy. Achieving adequate smoothing is a fundamental goal of these image smoothing algorithms. We show that partially ""erasing"" the appearance preservation facilitate adequate image smoothing. In this paper, we call this manipulation as Erasing Appearance Preservation (EAP). We conduct an user study, allowing users to indicate the ""erasing"" positions by drawing scribbles interactively, to verify the correctness and effectiveness of EAP. We observe the characteristics of human-indicated ""erasing"" positions, and then formulate a simple and effective 0-1 knapsack to automatically synthesize the ""erasing"" positions. We test our synthesized erasing positions in a majority of image smoothing methods. Experimental results and large-scale perceptual human judgments show that the EAP solution tends to encourage the pattern separation or elimination capabilities of image smoothing algorithms. We further study the performance of the EAP solution in many image decomposition problems to decompose textures, shadows, and the challenging specular reflections. We also present examinations of diversiform image manipulation applications like texture removal, retexturing, intrinsic decomposition, layer extraction, recoloring, material manipulation, etc. Due to the widespread applicability of image smoothing, the EAP is also likely to be used in more image editing applications."

Recent research efforts enable study for natural language grounded navigation in photo-realistic environments, e.g., following natural language instructions or dialog. However, existing methods tend to overfit training data in seen environments and fail to generalize well in previously unseen environments. To close the gap between seen and unseen environments, we aim at learning a generalized navigation model from two novel perspectives:(1) we introduce a multitask navigation model that can be seamlessly trained on both Vision-Language Navigation (VLN) and Navigation from Dialog History (NDH) tasks, which benefits from richer natural language guidance and effectively transfers knowledge across tasks;(2) we propose to learn environment-agnostic representations for the navigation policy that are invariant among the environments seen during training, thus generalizing better on unseen environments. Extensive experiments show that environment-agnostic multitask learning significantly reduces the performance gap between seen and unseen environments, and the navigation agent trained so outperforms baselines on unseen environments by 16\% (relative measure on success rate) on VLN and 120\% (goal progress) on NDH. Our submission to the CVDN leaderboard establishes a new state-of-the-art for the NDH task on the holdout test set. Code is available at https://github.com/google-research/valan ."

We propose a novel EMin framework for event-based vision model estimation. The framework extends previous event-based motion compensation algorithms to handle models whose outputs have arbitrary dimensions. The main motivation comes from estimating motion from events directly in 3D space (e. g. events augmented with depth), without projecting them onto an image plane. This is achieved by modelling the event alignment according to candidate parameters and minimising the resultant dispersion. We provide a family of suitable entropy loss functions and an efficient approximation whose complexity is only linear with the number of events (e. g. the complexity does not depend on the number of image pixels). The framework is evaluated on several motion estimation problems, including optical flow and rotational motion. As proof of concept, we also test our framework on 6-DOF estimation by performing the optimisation directly in 3D space."

Existing arts have shown promising efforts to deal with the blind deblurring task. However, most of the recent works assume the additive noise involved in the blurring process to be simple-distributed (i.e. Gaussian or Laplacian), while the real-world case is proved to be much more complicated. In this paper, we develop a new term to better fit the complex natural noise. Specifically, we use a weighted combination of a dense function (i.e. l2) and a newly designed enhanced sparse model termed as le, which is developed from two sparse models (i.e. l1 and l0), to fulfill the task. Moreover, we further suggest using le to regularize image gradients. Compared to the widely-adopted l0 sparse term, le can penalize more insignificant image details (Fig. 1). Based on the half-quadratic splitting method, we provide an effective scheme to optimize the overall formulation. Comprehensive evaluations on public datasets and real-world images demonstrate the superiority of the proposed method against state-of-the-art methods in terms of both speed and accuracy."

While deep learning-based classification is generally tackled using standardized approaches, a wide variety of techniques are employed for regression. In computer vision, one particularly popular such technique is that of confidence-based regression, which entails predicting a confidence value for each input-target pair $(x, y)$. While this approach has demonstrated impressive results, it requires important task-dependent design choices, and the predicted confidences lack a natural probabilistic meaning. We address these issues by proposing a general and conceptually simple regression method with a clear probabilistic interpretation. In our proposed approach, we create an energy-based model of the conditional target density $p(y | x)$, using a deep neural network to predict the un-normalized density from $(x, y)$. This model of $p(y | x)$ is trained by directly minimizing the associated negative log-likelihood, approximated using Monte Carlo sampling. We perform comprehensive experiments on four computer vision regression tasks. Our approach outperforms direct regression, as well as other probabilistic and confidence-based methods. Notably, our model achieves a $2.2\%$ AP improvement over Faster-RCNN for object detection on the COCO dataset, and sets a new state-of-the-art on visual tracking when applied for bounding box estimation. In contrast to confidence-based methods, our approach is also shown to be directly applicable to more general tasks such as age and head-pose estimation. Code is available at https://github.com/fregu856/ebms_regression."

Non-blind image deblurring is typically formulated as a linear least-squares problem regularized by natural priors on the corresponding sharp picture's gradients, which can be solved, for example, using a half-quadratic splitting method with Richardson fixed-point iterations for its least-squares updates and a proximal operator for the auxiliary variable updates. We propose to precondition the Richardson solver using approximate inverse filters of the (known) blur and natural image prior kernels. Using convolutions instead of a generic linear preconditioner allows extremely efficient parameter sharing across the image, and leads to significant gains in accuracy and/or speed compared to classical FFT and conjugate-gradient methods. More importantly, the proposed architecture is easily adapted to learning both the preconditioner and the proximal operator using CNN embeddings. This yields a simple and efficient algorithm for non-blind image deblurring which is fully interpretable, can be learned end to end, and whose accuracy matches or exceeds the state of the art, quite significantly, in the non-uniform case."

As an important computer vision task, 3d human pose estimation in a multi-camera, multi-person setting has received widespread attention and many interesting applications have been derived from it. Traditional approaches use a 3d pictorial structure model to handle this task. However, these models suffer from high computation costs and result in low accuracy in joint detection. Recently, especially since the introduction of Deep Neural Networks, one popular approach is to build a pipeline that involves three separate steps: (1) 2d skeleton detection in each camera view, (2) identification of matched 2d skeletons and (3) estimation of the 3d poses. Many existing works operate by feeding the 2d images and camera parameters through the three modules in a cascade fashion. However, all three operations can be highly correlated. For example, the 3d generation results may affect the results of detection in step 1, as does the matching algorithm in step 2. To address this phenomenon, we propose a novel end-to-end training scheme that brings the three separate modules into a single model. However, one outstanding problem of doing so is that the matching algorithm in step 2 appears to disjoint the pipeline. Therefore, we take our inspiration from the recent success in Capsule Networks, in which its Dynamic Routing step is also disjointed, but plays a crucial role in deciding how gradients are flowed from the upper to the lower layers. Similarly, a dynamic matching module in our work also decides the paths in which gradients flow from step 3 to step 1. Furthermore, as a large number of cameras are present, the existing matching algorithm either fails to deliver a robust performance or can be very inefficient. Thus, we additionally propose a novel matching algorithm that can match 2d poses from multiple views efficiently. The algorithm is robust and able to deal with situations of incomplete and false 2d detection as well."

The automated segmentation of buildings in remote sensing imagery is a challenging task that requires the accurate delineation of multiple building instances over typically large image areas. Manual methods are often laborious and current deep-learning-based approaches fail to delineate all building instances and do so with adequate accuracy. As a solution, we present Trainable Deep Active Contours (TDACs), an automatic image segmentation framework that intimately unites Convolutional Neural Networks (CNNs) and Active Contour Models (ACMs). The Eulerian energy functional of the ACM component includes per-pixel parameter maps that are predicted by the backbone CNN, which also initializes the ACM. Importantly, both the ACM and CNN components are fully implemented in TensorFlow and the entire TDAC architecture is end-to-end automatically differentiable and backpropagation trainable without user intervention. TDAC yields fast, accurate, and fully automatic simultaneous delineation of arbitrarily many buildings in the image. We validate the model on two publicly available aerial image datasets for building segmentation, and our results demonstrate that TDAC establishes a new state-of-the-art performance."

We present a new method that views object detection as a direct set prediction. Our approach streamlines the detection pipeline, effectively removing the need for many hand-designed components like a non-maximum suppression procedure or anchor generation that explicitly encode our prior knowledge about the task. The main ingredients of the new framework, called DEtection TRansformer or DETR, are a set-based global loss that forces unique predictions via bipartite matching, and a transformer encoder-decoder architecture. Given a fixed small set of learned object queries, DETR reasons about the relations of the objects and the global image context to directly output the final set of predictions in parallel. The new model is conceptually simple and does not require a specialized library, unlike many other modern detectors. DETR demonstrates accuracy and run-time performance on par with the well-established and highly-optimized Faster R-CNN baseline on the challenging COCO object detection dataset. Moreover, we show that DETR can be easily generalized to produce a competitive panoptic segmentation prediction in a unified manner."

Coded aperture snapshot spectral imaging (CASSI) is an effective tool to capture real-world 3D hyperspectral images. While a number of existing work has been conducted for hardware and algorithm design, we make a step towards the low-cost solution that enjoys video-rate high-quality reconstruction. To make solid progress on this challenging yet under-investigated task, we reproduce a stable single disperser (SD) CASSI system to gather large-scale real-world CASSI data and propose a novel deep convolutional network to carry out the real-time reconstruction by using self-attention. In order to jointly capture the self-attention across different dimensions in hyperspectral images (i.e., channel-wise spectral correlation and non-local spatial regions), we propose Spatial-Spectral Self-Attention (TSA) to process each dimension sequentially, yet in an order-independent manner. We employ TSA in an encoder-decoder network, dubbed TSA-Net, to reconstruct the desired 3D cube. Furthermore, we investigate how noise affects the results and propose to add shot noise in model training, which improves the real data results significantly. We hope our large-scale CASSI data serve as a benchmark in future research and our TSA model as a baseline in deep learning based reconstruction algorithms."

Anomaly detection in retinal image refers to the identification of abnormality caused by various retinal diseases/lesions, by only leveraging normal images in training phase. Normal images from healthy subjects often have regular structures (e.g., the structured blood vessels in the fundus image, or structured anatomy in optical coherence tomography image). On the contrary, the diseases and lesions often destroy these structures. Motivated by this, we propose to leverage the relation between the image texture and structure to design a deep neural network for anomaly detection. Specifically, we first extract the structure of the retinal images, then we combine both the structure features and the last layer features extracted from original health image to reconstruct the original input healthy image. The image feature provides the texture information and guarantees the uniqueness of the image recovered from the structure. In the end, we further utilize the reconstructed image to extract the structure and measure the difference between structure extracted from original and the reconstructed image. On the one hand, minimizing the reconstruction difference behaves like a regularizer to guarantee that the image is corrected reconstructed. On the other hand, such structure difference can also be used as a metric for normality measurement. The whole network is termed as P-Net because it has a ""P"" shape. Extensive experiments on RESC dataset and iSee dataset validate the effectiveness of our approach for anomaly detection in retinal images. Further, our method also generalizes well to novel class discovery for retinal images and anomaly detection for real-world images. "

Weakly supervised object detection (WSOD) has attracted extensive research attention due to its great flexibility of exploiting large-scale image-level annotation for detector training. Whilst deep residual networks such as ResNet and DenseNet have become the standard backbones for many computer vision tasks, the cutting-edge WSOD methods still rely on plain networks, e.g., VGG, as backbones. It is indeed not trivial to employ deep residual networks for WSOD, which even shows significant deterioration of detection accuracy and non-convergence. In this paper, we discover the intrinsic root with sophisticated analysis and propose a sequence of design principles to take full advantages of deep residual learning for WSOD from the perspectives of adding redundancy, improving robustness and aligning features. First, a redundant adaptation neck is key for effective object instance localization and discriminative feature learning. Second, small-kernel convolutions and MaxPool down-samplings help improve the robustness of information flow, which gives finer object boundaries and make the detector more sensitivity to small objects. Third, dilated convolution is essential to align the proposal features and exploit diverse local information by extracting high-resolution feature maps. Extensive experiments show that the proposed principles enable deep residual networks to establishes new state-of-the-arts on PASCAL VOC and MS COCO."

This paper presents a novel relational network for group activity recognition. The core of our network is to augment the conditional random fields (CRF), amenable to learning inter-dependency of correlated observations, with the newly devised temporal and spatial self-attention to learn the temporal evolution and spatial relational contexts of every actor in videos. Such a combination utilizes the global receptive fields of self-attention to construct a spatio-temporal graph topology to address the temporal dependency and non-local relationships of the actors. The network first uses the temporal self-attention along with the spatial self-attention, which considers multiple cliques with different scales of locality to account for the diversity of the actors' relationships in group activities, to model the pairwise energy of CRF. Afterward, to accommodate the distinct characteristics of each video, a new mean-field inference algorithm with dynamic halting is also addressed. Finally, a bidirectional universal transformer encoder (UTE), which combines both of the forward and backward temporal context information, is used to aggregate the relational contexts and scene information for group activity recognition. Simulations show that the proposed approach surpasses the state-of-the-art methods on the widespread Volleyball and Collective Activity datasets."