CVPR 2020 论文列表

Deep learning (DL) methods have achieved state-of-the-art performance in the task of single image rain removal. Most of current DL architectures, however, are still lack of sufficient interpretability and not fully integrated with physical structures inside general rain streaks. To this issue, in this paper, we propose a model-driven deep neural network for the task, with fully interpretable network structures. Specifically, based on the convolutional dictionary learning mechanism for representing rain, we propose a novel single image deraining model and utilize the proximal gradient descent technique to design an iterative algorithm only containing simple operators for solving the model. Such a simple implementation scheme facilitates us to unfold it into a new deep network architecture, called rain convolutional dictionary network (RCDNet), with almost every network module one-to-one corresponding to each operation involved in the algorithm. By end-to-end training the proposed RCDNet, all the rain kernels and proximal operators can be automatically extracted, faithfully characterizing the features of both rain and clean background layers, and thus naturally lead to its better deraining performance, especially in real scenarios. Comprehensive experiments substantiate the superiority of the proposed network, especially its well generality to diverse testing scenarios and good interpretability for all its modules, as compared with state-of-the-arts both visually and quantitatively.

Scene, as the crucial unit of storytelling in movies, contains complex activities of actors and their interactions in a physical environment. Identifying the composition of scenes serves as a critical step towards semantic understanding of movies. This is very challenging - compared to the videos studied in conventional vision problems, e.g. action recognition, as scenes in movies usually contain much richer temporal structures and more complex semantic information. Towards this goal, we scale up the scene segmentation task by building a large-scale video dataset MovieScenes, which contains 21K annotated scene segments from 150 movies. We further propose a local-to-global scene segmentation framework, which integrates multi-modal information across three levels, i.e. clip, segment, and movie. This framework is able to distill complex semantics from hierarchical temporal structures over a long movie, providing top-down guidance for scene segmentation. Our experiments show that the proposed network is able to segment a movie into scenes with high accuracy, consistently outperforming previous methods. We also found that pretraining on our MovieScenes can bring significant improvements to the existing approaches.

Under the conventional diffuse shading model with unknown directional lighting, the set of quadratic surface shapes that are consistent with the spatial derivatives of intensity at a single image point is a two-dimensional algebraic variety embedded in the five-dimensional space of quadratic shapes. We describe the geometry of this variety, and we introduce a concise feedforward model that computes an explicit, differentiable approximation of the variety from the intensity and its derivatives at any single image point. The result is a parallelizable processor that operates at each image point and produces a lighting-invariant descriptor of the continuous set of compatible surface shapes at the point. We describe two applications of this processor: two-shot uncalibrated photometric stereo and quadratic-surface shape from shading.

3D object detection on point clouds finds many applications. However, most known point cloud object detection methods did not adequately accommodate the characteristics (e.g., sparsity) of point clouds, and thus some key semantic information (e.g., shape information) is not well captured. In this paper, we propose a new graph convolution (GConv) based hierarchical graph network (HGNet) for 3D object detection, which processes raw point clouds directly to predict 3D bounding boxes. HGNet effectively captures the relationship of the points and utilizes the multi-level semantics for object detection. Specially, we propose a novel shape-attentive GConv (SA-GConv) to capture the local shape features, by modelling the relative geometric positions of points to describe object shapes. An SA-GConv based U-shape network captures the multi-level features, which are mapped into an identical feature space by an improved voting module and then further utilized to generate proposals. Next, a new GConv based Proposal Reasoning Module reasons on the proposals considering the global scene semantics, and the bounding boxes are then predicted. Consequently, our new framework outperforms state-of-the-art methods on two large-scale point cloud datasets, by 4% mean average precision (mAP) on SUN RGB-D and by 3% mAP on ScanNet-V2.

Due to the highly non-convex nature of large-scale robust parameter estimation, avoiding poor local minima is challenging in real-world applications where input data is contaminated by a large or unknown fraction of outliers. In this paper, we introduce a novel solver for robust estimation that possesses a strong ability to escape poor local minima. Our algorithm is built upon the class of traditional graduated optimization techniques, which are considered state-of-the-art local methods to solve problems having many poor minima. The novelty of our work lies in the introduction of an adaptive kernel (or residual) scaling scheme, which allows us to achieve faster convergence rates. Like other existing methods that aim to return good local minima for robust estimation tasks, our method relaxes the original robust problem, but adapts a filter framework from non-linear constrained optimization to automatically choose the level of relaxation. Experimental results on real large-scale datasets such as bundle adjustment instances demonstrate that our proposed method achieves competitive results.

Neural networks have greatly boosted performance in computer vision by learning powerful representations of input data. The drawback of end-to-end training for maximal overall performance are black-box models whose hidden representations are lacking interpretability: Since distributed coding is optimal for latent layers to improve their robustness, attributing meaning to parts of a hidden feature vector or to individual neurons is hindered. We formulate interpretation as a translation of hidden representations onto semantic concepts that are comprehensible to the user. The mapping between both domains has to be bijective so that semantic modifications in the target domain correctly alter the original representation. The proposed invertible interpretation network can be transparently applied on top of existing architectures with no need to modify or retrain them. Consequently, we translate an original representation to an equivalent yet interpretable one and backwards without affecting the expressiveness and performance of the original. The invertible interpretation network disentangles the hidden representation into separate, semantically meaningful concepts. Moreover, we present an efficient approach to define semantic concepts by only sketching two images and also an unsupervised strategy. Experimental evaluation demonstrates the wide applicability to interpretation of existing classification and image generation networks as well as to semantically guided image manipulation.

In video understanding, action spotting consists in temporally localizing human-induced events annotated with single timestamps. In this paper, we propose a novel loss function that specifically considers the temporal context naturally present around each action, rather than focusing on the single annotated frame to spot. We benchmark our loss on a large dataset of soccer videos, SoccerNet, and achieve an improvement of 12.8% over the baseline. We show the generalization capability of our loss for generic activity proposals and detection on ActivityNet, by spotting the beginning and the end of each activity. Furthermore, we provide an extended ablation study and display challenging cases for action spotting in soccer videos. Finally, we qualitatively illustrate how our loss induces a precise temporal understanding of actions and show how such semantic knowledge can be used for automatic highlights generation.

Implicit Generative Models (IGMs) such as GANs have emerged as effective data-driven models for generating samples, particularly images. In this paper, we formulate the problem of learning an IGM as minimizing the expected distance between characteristic functions. Specifically, we minimize the distance between characteristic functions of the real and generated data distributions under a suitably-chosen weighting distribution. This distance metric, which we term as the characteristic function distance (CFD), can be (approximately) computed with linear time-complexity in the number of samples, in contrast with the quadratic-time Maximum Mean Discrepancy (MMD). By replacing the discrepancy measure in the critic of a GAN with the CFD, we obtain a model that is simple to implement and stable to train. The proposed metric enjoys desirable theoretical properties including continuity and differentiability with respect to generator parameters, and continuity in the weak topology. We further propose a variation of the CFD in which the weighting distribution parameters are also optimized during training; this obviates the need for manual tuning, and leads to an improvement in test power relative to CFD. We demonstrate experimentally that our proposed method outperforms WGAN and MMD-GAN variants on a variety of unsupervised image generation benchmarks.

We present a convex optimization approach for generalized essential matrix (GEM) estimation. The six-point minimal solver for the GEM has poor numerical stability and applies only for a minimal number of points. Existing non-minimal solvers for GEM estimation rely on either local optimization or relinearization techniques, which impedes high accuracy in common scenarios. Our proposed non-minimal solver minimizes the sum of squared residuals by reformulating the problem as a quadratically constrained quadratic program. The globally optimal solution is thus obtained by a semidefinite relaxation. The algorithm retrieves certifiably globally optimal solutions to the original non-convex problem in polynomial time. We also provide the necessary and sufficient conditions to recover the optimal GEM from the relaxed problems. The improved performance is demonstrated over experiments on both synthetic and real multi-camera systems.

We present a data-driven approach for 4D space-time visualization of dynamic events from videos captured by hand-held multiple cameras. Key to our approach is the use of self-supervised neural networks specific to the scene to compose static and dynamic aspects of an event. Though captured from discrete viewpoints, this model enables us to move around the space-time of the event continuously. This model allows us to create virtual cameras that facilitate: (1) freezing the time and exploring views; (2) freezing a view and moving through time; and (3) simultaneously changing both time and view. We can also edit the videos and reveal occluded objects for a given view if it is visible in any of the other views. We validate our approach on challenging in-the-wild events captured using up to 15 mobile cameras.

his paper contributes a novel realtime multi-person motion capture algorithm using multiview video inputs. Due to the heavy occlusions and closely interacting motions in each view, joint optimization on the multiview images and multiple temporal frames is indispensable, which brings up the essential challenge of realtime efficiency. To this end, for the first time, we unify per-view parsing, cross-view matching, and temporal tracking into a single optimization framework, i.e., a 4D association graph that each dimension (image space, viewpoint and time) can be treated equally and simultaneously. To solve the 4D association graph efficiently, we further contribute the idea of 4D limb bundle parsing based on heuristic searching, followed with limb bundle assembling by proposing a bundle Kruskal's algorithm. Our method enables a realtime motion capture system running at 30fps using 5 cameras on a 5-person scene. Benefiting from the unified parsing, matching and tracking constraints, our method is robust to noisy detection due to severe occlusions and close interacting motions, and achieves high-quality online pose reconstruction quality. The proposed method outperforms state-of-the-art methods quantitatively without using high-level appearance information.

Current supervised methods for facial landmark detection require a large amount of training data and may suffer from overfitting to specific datasets due to the massive number of parameters. We introduce a semi-supervised method in which the crucial idea is to first generate implicit face knowledge from the large amounts of unlabeled images of faces available today. In a first, completely unsupervised stage, we train an adversarial autoencoder to reconstruct faces via a low-dimensional face embedding. In a second, supervised stage, we interleave the decoder with transfer layers to retask the generation of color images to the prediction of landmark heatmaps. Our framework (3FabRec) achieves state-of-the-art performance on several common benchmarks and, most importantly, is able to maintain impressive accuracy on extremely small training sets down to as few as 10 images. As the interleaved layers only add a low amount of parameters to the decoder, inference runs at several hundred FPS on a GPU.

For depth-based 3D action recognition, one essential issue is to represent 3D motion pattern effectively and efficiently. To this end, 3D dynamic voxel (3DV) is proposed as a novel 3D motion representation manner. With 3D space voxelization, the key idea of 3DV is to encode the 3D motion information within depth video into a regular voxel set (i.e., 3DV) compactly, via temporal rank pooling. Each available 3DV voxel intrinsically involves 3D spatial and motion feature for 3D action description. 3DV is then abstracted as a point set and input into PointNet++ for 3D action recognition, in the end-to-end learning way. The intuition for transferring 3DV into the point set form is that, PointNet++ is lightweight and effective for deep feature learning towards point set. Since 3DV may loose appearance clue, a multi-stream 3D action recognition manner is also proposed to learn motion and appearance feature jointly. To extract richer temporal order information of actions, we also split the depth video into temporal segments and encode this procedure in 3DV integrally. The extensive experiments on the well-established benchmark datasets (e.g., NTU RGB+D 120 and NTU RGB+D 60) demonstrate the superiority of our proposition. Impressively, we acquire the accuracy of 82.4% and 93.5% on NTU RGB+D 120 with the cross-subject and cross-setup test setting respectively. 3DV's code is available at https://github.com/3huo/3DV-Action.

Prevalence of voxel-based 3D single-stage detectors contrast with underexplored point-based methods. In this paper, we present a lightweight point-based 3D single stage object detector 3DSSD to achieve decent balance of accuracy and efficiency. In this paradigm, all upsampling layers and the refinement stage, which are indispensable in all existing point-based methods, are abandoned. We instead propose a fusion sampling strategy in downsampling process to make detection on less representative points feasible. A delicate box prediction network, including a candidate generation layer and an anchor-free regression head with a 3D center-ness assignment strategy, is developed to meet the demand of high accuracy and speed. Our 3DSSD paradigm is an elegant single-stage anchor-free one. We evaluate it on widely used KITTI dataset and more challenging nuScenes dataset. Our method outperforms all state-of-the-art voxel-based single-stage methods by a large margin, and even yields comparable performance with two-stage point-based methods, with amazing inference speed of 25+ FPS, 2x faster than former state-of-the-art point-based methods.

We present 3DRegNet, a novel deep learning architecture for the registration of 3D scans. Given a set of 3D point correspondences, we build a deep neural network to address the following two challenges: (i) classification of the point correspondences into inliers/outliers, and (ii) regression of the motion parameters that align the scans into a common reference frame. With regard to regression, we present two alternative approaches: (i) a Deep Neural Network (DNN) registration and (ii) a Procrustes approach using SVD to estimate the transformation. Our correspondence-based approach achieves a higher speedup compared to competing baselines. We further propose the use of a refinement network, which consists of a smaller 3DRegNet as a refinement to improve the accuracy of the registration. Extensive experiments on two challenging datasets demonstrate that we outperform other methods and achieve state-of-the-art results. The code is available.

In this work we present a novel publicly available stereo based 3D RGB dataset for multi-object zebrafish tracking, called 3D-ZeF. Zebrafish is an increasingly popular model organism used for studying neurological disorders, drug addiction, and more. Behavioral analysis is often a critical part of such research. However, visual similarity, occlusion, and erratic movement of the zebrafish makes robust 3D tracking a challenging and unsolved problem. The proposed dataset consists of eight sequences with a duration between 15-120 seconds and 1-10 free moving zebrafish. The videos have been annotated with a total of 86,400 points and bounding boxes. Furthermore, we present a complexity score and a novel open-source modular baseline system for 3D tracking of zebrafish. The performance of the system is measured with respect to two detectors: a naive approach and a Faster R-CNN based fish head detector. The system reaches a MOTA of up to 77.6%. Links to the code and dataset is available at the project page http://vap.aau.dk/3d-zef

We present 3D-MPA, a method for instance segmentation on 3D point clouds. Given an input point cloud, we propose an object-centric approach where each point votes for its object center. We sample object proposals from the predicted object centers. Then, we learn proposal features from grouped point features that voted for the same object center. A graph convolutional network introduces inter-proposal relations, providing higher-level feature learning in addition to the lower-level point features. Each proposal comprises a semantic label, a set of associated points over which we define a foreground-background mask, an objectness score and aggregation features. Previous works usually perform non-maximum-suppression (NMS) over proposals to obtain the final object detections or semantic instances. However, NMS can discard potentially correct predictions. Instead, our approach keeps all proposals and groups them together based on the learned aggregation features. We show that grouping proposals improves over NMS and outperforms previous state-of-the-art methods on the tasks of 3D object detection and semantic instance segmentation on the ScanNetV2 benchmark and the S3DIS dataset.

The goal of the Semantic Scene Completion (SSC) task is to simultaneously predict a completed 3D voxel representation of volumetric occupancy and semantic labels of objects in the scene from a single-view observation. Since the computational cost generally increases explosively along with the growth of voxel resolution, most current state-of-the-arts have to tailor their framework into a low-resolution representation with the sacrifice of detail prediction. Thus, voxel resolution becomes one of the crucial difficulties that lead to the performance bottleneck. In this paper, we propose to devise a new geometry-based strategy to embed depth information with low-resolution voxel representation, which could still be able to encode sufficient geometric information, e.g., room layout, object's sizes and shapes, to infer the invisible areas of the scene with well structure-preserving details. To this end, we first propose a novel 3D sketch-aware feature embedding to explicitly encode geometric information effectively and efficiently. With the 3D sketch in hand, we further devise a simple yet effective semantic scene completion framework that incorporates a light-weight 3D Sketch Hallucination module to guide the inference of occupancy and the semantic labels via a semi-supervised structure prior learning strategy. We demonstrate that our proposed geometric embedding works better than the depth feature learning from habitual SSC frameworks. Our final model surpasses state- of-the-arts consistently on three public benchmarks, which only requires 3D volumes of 60 x 36 x 60 resolution for both input and output.

We propose a method for converting a single RGB-D input image into a 3D photo, i.e., a multi-layer representation for novel view synthesis that contains hallucinated color and depth structures in regions occluded in the original view. We use a Layered Depth Image with explicit pixel connectivity as underlying representation, and present a learning-based inpainting model that iteratively synthesizes new local color-and-depth content into the occluded region in a spatial context-aware manner. The resulting 3D photos can be efficiently rendered with motion parallax using standard graphics engines. We validate the effectiveness of our method on a wide range of challenging everyday scenes and show less artifacts when compared with the state-of-the-arts.