CVPR 2020 论文列表

Deep neural networks have improved image classification dramatically over the past decade, but have done so by focusing on performance measures that treat all classes other than the ground truth as equally wrong. This has led to a situation in which mistakes are less likely to be made than before, but are equally likely to be absurd or catastrophic when they do occur. Past works have recognised and tried to address this issue of mistake severity, often by using graph distances in class hierarchies, but this has largely been neglected since the advent of the current deep learning era in computer vision. In this paper, we aim to renew interest in this problem by reviewing past approaches and proposing two simple methods which outperform the prior art under several metrics on two large datasets with complex class hierarchies: tieredImageNet and iNaturalist'19.

Deep neural networks (DNNs) have been applied in class incremental learning, which aims to solve common real-world problems of learning new classes continually. One drawback of standard DNNs is that they are prone to catastrophic forgetting. Knowledge distillation (KD) is a commonly used technique to alleviate this problem. In this paper, we demonstrate it can indeed help the model to output more discriminative results within old classes. However, it cannot alleviate the problem that the model tends to classify objects into new classes, causing the positive effect of KD to be hidden and limited. We observed that an important factor causing catastrophic forgetting is that the weights in the last fully connected (FC) layer are highly biased in class incremental learning. In this paper, we propose a simple and effective solution motivated by the aforementioned observations to address catastrophic forgetting. Firstly, we utilize KD to maintain the discrimination within old classes. Then, to further maintain the fairness between old classes and new classes, we propose Weight Aligning (WA) that corrects the biased weights in the FC layer after normal training process. Unlike previous work, WA does not require any extra parameters or a validation set in advance, as it utilizes the information provided by the biased weights themselves. The proposed method is evaluated on ImageNet-1000, ImageNet-100, and CIFAR-100 under various settings. Experimental results show that the proposed method can effectively alleviate catastrophic forgetting and significantly outperform state-of-the-art methods.

Convolutional neural networks have witnessed remarkable improvements in computational efficiency in recent years. A key driving force has been the idea of trading-off model expressivity and efficiency through a combination of 1x1 and depth-wise separable convolutions in lieu of a standard convolutional layer. The price of the efficiency, however, is the sub-optimal flow of information across space and channels in the network. To overcome this limitation, we present MUXConv, a layer that is designed to increase the flow of information by progressively multiplexing channel and spatial information in the network, while mitigating computational complexity. Furthermore, to demonstrate the effectiveness of MUXConv, we integrate it within an efficient multi-objective evolutionary algorithm to search for the optimal model hyper-parameters while simultaneously optimizing accuracy, compactness, and computational efficiency. On ImageNet, the resulting models, dubbed MUXNets, match the performance (75.3% top-1 accuracy) and multiply-add operations (218M) of MobileNetV3 while being 1.6x more compact, and outperform other mobile models in all the three criteria. MUXNet also performs well under transfer learning and when adapted to object detection. On the ChestX-Ray 14 benchmark, its accuracy is comparable to the state-of-the-art while being 3.3x more compact and 14x more efficient. Similarly, detection on PASCAL VOC 2007 is 1.2% more accurate, 28% faster and 6% more compact compared to MobileNetV2.

We propose to incorporate neural architecture search (NAS) into general-purpose multi-task learning (GP-MTL). Existing NAS methods typically define different search spaces according to different tasks. In order to adapt to different task combinations (i.e., task sets), we disentangle the GP-MTL networks into single-task backbones (optionally encode the task priors), and a hierarchical and layerwise features sharing/fusing scheme across them. This enables us to design a novel and general task-agnostic search space, which inserts cross-task edges (i.e., feature fusion connections) into fixed single-task network backbones. Moreover, we also propose a novel single-shot gradient-based search algorithm that closes the performance gap between the searched architectures and the final evaluation architecture. This is realized with a minimum entropy regularization on the architecture weights during the search phase, which makes the architecture weights converge to near-discrete values and therefore achieves a single model. As a result, our searched model can be directly used for evaluation without (re-)training from scratch. We perform extensive experiments using different single-task backbones on various task sets, demonstrating the promising performance obtained by exploiting the hierarchical and layerwise features, as well as the desirable generalizability to different i) task sets and ii) single-task backbones. The code of our paper is available at https://github.com/bhpfelix/MTLNAS.

We present MSeg, a composite dataset that unifies se- mantic segmentation datasets from different domains. A naive merge of the constituent datasets yields poor performance due to inconsistent taxonomies and annotation practices. We reconcile the taxonomies and bring the pixel-level annotations into alignment by relabeling more than 220,000 object masks in more than 80,000 images. The resulting composite dataset enables training a single semantic segmentation model that functions effectively across domains and generalizes to datasets that were not seen during training. We adopt zero-shot cross-dataset transfer as a benchmark to systematically evaluate a model's robustness and show that MSeg training yields substantially more robust models in comparison to training on individual datasets or naive mixing of datasets without the presented contributions. A model trained on MSeg ranks first on the WildDash leaderboard for robust semantic segmentation, with no exposure to WildDash data during training.

While Generative Adversarial Networks (GANs) have seen huge successes in image synthesis tasks, they are notoriously difficult to adapt to different datasets, in part due to instability during training and sensitivity to hyperparameters. One commonly accepted reason for this instability is that gradients passing from the discriminator to the generator become uninformative when there isn't enough overlap in the supports of the real and fake distributions. In this work, we propose the Multi-Scale Gradient Generative Adversarial Network (MSG-GAN), a simple but effective technique for addressing this by allowing the flow of gradients from the discriminator to the generator at multiple scales. This technique provides a stable approach for high resolution image synthesis, and serves as an alternative to the commonly used progressive growing technique. We show that MSG-GAN converges stably on a variety of image datasets of different sizes, resolutions and domains, as well as different types of loss functions and architectures, all with the same set of fixed hyperparameters. When compared to state-of-the-art GANs, our approach matches or exceeds the performance in most of the cases we tried.

Conventional cell tracking methods detect multiple cells in each frame (detection) and then associate the detection results in successive time-frames (association). Most cell tracking methods perform the association task independently from the detection task. However, there is no guarantee of preserving coherence between these tasks, and lack of coherence may adversely affect tracking performance. In this paper, we propose the Motion and Position Map (MPM) that jointly represents both detection and association for not only migration but also cell division. It guarantees coherence such that if a cell is detected, the corresponding motion flow can always be obtained. It is a simple but powerful method for multi-object tracking in dense environments. We compared the proposed method with current tracking methods under various conditions in real biological images and found that it outperformed the state-of-the-art (+5.2% improvement compared to the second-best).

In late fusion, each modality is processed in a separate unimodal Convolutional Neural Network (CNN) stream and the scores of each modality are fused at the end. Due to its simplicity, late fusion is still the predominant approach in many state-of-the-art multimodal applications. In this paper, we present a simple neural network module for leveraging the knowledge from multiple modalities in convolutional neural networks. The proposed unit, named Multimodal Transfer Module (MMTM), can be added at different levels of the feature hierarchy, enabling slow modality fusion. Using squeeze and excitation operations, MMTM utilizes the knowledge of multiple modalities to recalibrate the channel-wise features in each CNN stream. Unlike other intermediate fusion methods, the proposed module could be used for feature modality fusion in convolution layers with different spatial dimensions. Another advantage of the proposed method is that it could be added among unimodal branches with minimum changes in the their network architectures, allowing each branch to be initialized with existing pretrained weights. Experimental results show that our framework improves the recognition accuracy of well-known multimodal networks. We demonstrate state-of-the-art or competitive performance on four datasets that span the task domains of dynamic hand gesture recognition, speech enhancement, and action recognition with RGB and body joints.

In this paper, we address the 3D object detection task by capturing multi-level contextual information with the self-attention mechanism and multi-scale feature fusion. Most existing 3D object detection methods recognize objects individually, without giving any consideration on contextual information between these objects. Comparatively, we propose Multi-Level Context VoteNet (MLCVNet) to recognize 3D objects correlatively, building on the state-of-the-art VoteNet. We introduce three context modules into the voting and classifying stages of VoteNet to encode contextual information at different levels. Specifically, a Patch-to-Patch Context (PPC) module is employed to capture contextual information between the point patches, before voting for their corresponding object centroid points. Subsequently, an Object-to-Object Context (OOC) module is incorporated before the proposal and classification stage, to capture the contextual information between object candidates. Finally, a Global Scene Context (GSC) module is designed to learn the global scene context. We demonstrate these by capturing contextual information at patch, object and scene levels. Our method is an effective way to promote detection accuracy, achieving new state-of-the-art detection performance on challenging 3D object detection datasets, i.e., SUN RGBD and ScanNet. We also release our code at https://github.com/NUAAXQ/MLCVNet.

In this paper, we explore synthesizing person images with multiple conditions for various backgrounds. To this end, we propose a framework named "MISC" for conditional image generation and image compositing. For conditional image generation, we improve the existing condition injection mechanisms by leveraging the inter-condition correlations. For the image compositing, we theoretically prove the weaknesses of the cutting-edge methods, and make it more robust by removing the spatially-invariance constraint, and enabling the bounding mechanism and the spatial adaptability. We show the effectiveness of our method on the Video Instance-level Parsing dataset, and demonstrate the robustness through controllability tests.

We present a convex mixed-integer programming formulation for non-rigid shape matching. To this end, we propose a novel shape deformation model based on an efficient low-dimensional discrete model, so that finding a globally optimal solution is tractable in (most) practical cases. Our approach combines several favourable properties, namely it is independent of the initialisation, it is much more efficient to solve to global optimality compared to analogous quadratic assignment problem formulations, and it is highly flexible in terms of the variants of matching problems it can handle. Experimentally we demonstrate that our approach outperforms existing methods for sparse shape matching, that it can be used for initialising dense shape matching methods, and we showcase its flexibility on several examples.

Existing Temporal Activity Localization (TAL) methods largely adopt strong supervision for model training, which requires (1) vast amounts of untrimmed videos per each activity category and (2) accurate segment-level boundary annotations (start time and end time) for every instance. This poses a critical restriction to the current methods in practical scenarios where not only segment-level annotations are expensive to obtain, but many activity categories are also rare and unobserved during training. Therefore, Can we learn a TAL model under weak supervision that can localize unseen activity classes? To address this scenario, we define a novel example-based TAL problem called Minimum Effort Temporal Activity Localization (METAL): Given only a few examples, the goal is to find the occurrences of semantically-related segments in an untrimmed video sequence while model training is only supervised by the video-level annotation. Towards this objective, we propose a novel Similarity Pyramid Network (SPN) that adopts the few-shot learning technique of Relation Network and directly encodes hierarchical multi-scale correlations, which we learn by optimizing two complimentary loss functions in an end-to-end manner. We evaluate the SPN on the THUMOS'14 and ActivityNet datasets, of which we rearrange the videos to fit the METAL setup. Results show that our SPN achieves performance superior or competitive to state-of-the-art approaches with stronger supervision.

Body orientation estimation provides crucial visual cues in many applications, including robotics and autonomous driving. It is particularly desirable when 3-D pose estimation is difficult to infer due to poor image resolution, occlusion or indistinguishable body parts. We present COCO-MEBOW (Monocular Estimation of Body Orientation in the Wild), a new large-scale dataset for orientation estimation from a single in-the-wild image. The body-orientation labels for around 130K human bodies within 55K images from the COCO dataset have been collected using an efficient and high-precision annotation pipeline. We also validated the benefits of the dataset. First, we show that our dataset can substantially improve the performance and the robustness of a human body orientation estimation model, the development of which was previously limited by the scale and diversity of the available training data. Additionally, we present a novel triple-source solution for 3-D human pose estimation, where 3-D pose labels, 2-D pose labels, and our body-orientation labels are all used in joint training. Our model significantly outperforms state-of-the-art dual-source solutions for monocular 3-D human pose estimation, where training only uses 3-D pose labels and 2-D pose labels. This substantiates an important advantage of MEBOW for 3-D human pose estimation, which is particularly appealing because the per-instance labeling cost for body orientations is far less than that for 3-D poses. The work demonstrates high potential of MEBOW in addressing real-world challenges involving understanding human behaviors. Further information of this work is available at https://chenyanwu.github.io/MEBOW/.

Nowadays, driven by the increasing concern on diet and health, food computing has attracted enormous attention from both industry and research community. One of the most popular research topics in this domain is Food Retrieval, due to its profound influence on health-oriented applications. In this paper, we focus on the task of cross-modal retrieval between food images and cooking recipes. We present Modality-Consistent Embedding Network (MCEN) that learns modality-invariant representations by projecting images and texts to the same embedding space. To capture the latent alignments between modalities, we incorporate stochastic latent variables to explicitly exploit the interactions between textual and visual features. Importantly, our method learns the cross-modal alignments during training but computes embeddings of different modalities independently at inference time for the sake of efficiency. Extensive experimental results clearly demonstrate that the proposed MCEN outperforms all existing approaches on the benchmark Recipe1M dataset and requires less computational cost.

Recent interest in self-supervised dense tracking has yielded rapid progress, but performance still remains far from supervised methods. We propose a dense tracking model trained on videos without any annotations that surpasses previous self-supervised methods on existing benchmarks by a significant margin (+15%), and achieves performance comparable to supervised methods. In this paper, we first reassess the traditional choices used for self-supervised training and reconstruction loss by conducting thorough experiments that finally elucidate the optimal choices. Second, we further improve on existing methods by augmenting our architecture with a crucial memory component. Third, we benchmark on large-scale semi-supervised video object segmentation (aka. dense tracking), and propose a new metric: generalizability. Our first two contributions yield a self-supervised network that for the first time is competitive with supervised methods on standard evaluation metrics of dense tracking. When measuring generalizability, we show self-supervised approaches are actually superior to the majority of supervised methods. We believe this new generalizability metric can better capture the real-world use-cases for dense tracking, and will spur new interest in this research direction.

The ambiguity in image matching is one of main factors decreasing the quality of the 3D model reconstructed by PatchMatch based multiple view stereo. In this paper, we present a novel method, matching ambiguity reduced multiple view stereo (MARMVS) to address this issue. The MARMVS handles the ambiguity in image matching process with three newly proposed strategies: 1) The matching ambiguity is measured by the differential geometry property of image surface with epipolar constraint, which is used as a critical criterion for optimal scale selection of every single pixel with corresponding neighbouring images. 2) The depth of every pixel is initialized to be more close to the true depth by utilizing the depths of its surrounding sparse feature points, which yields faster convergency speed in the following PatchMatch stereo and alleviates the ambiguity introduced by self similar structures of the image. 3) In the last propagation of the PatchMatch stereo, higher priorities are given to those planes with the related 2D image patch possesses less ambiguity, this strategy further propagates a correctly reconstructed surface to raw texture regions. In addition, the proposed method is very efficient even running on consumer grade CPUs, due to proper parameterization and discretization in the depth map computation step. The MARMVS is validated on public benchmarks, and experimental results demonstrate competing performance against the state of the art.

Autonomous vehicles are expected to drive in complex scenarios with several independent non cooperating agents. Path planning for safely navigating in such environments can not just rely on perceiving present location and motion of other agents. It requires instead to predict such variables in a far enough future. In this paper we address the problem of multimodal trajectory prediction exploiting a Memory Augmented Neural Network. Our method learns past and future trajectory embeddings using recurrent neural networks and exploits an associative external memory to store and retrieve such embeddings. Trajectory prediction is then performed by decoding in-memory future encodings conditioned with the observed past. We incorporate scene knowledge in the decoding state by learning a CNN on top of semantic scene maps. Memory growth is limited by learning a writing controller based on the predictive capability of existing embeddings. We show that our method is able to natively perform multi-modal trajectory prediction obtaining state-of-the art results on three datasets. Moreover, thanks to the non-parametric nature of the memory module, we show how once trained our system can continuously improve by ingesting novel patterns.

We propose MAGSAC++ and Progressive NAPSAC sampler, P-NAPSAC in short. In MAGSAC++, we replace the model quality and polishing functions of the original method by an iteratively re-weighted least-squares fitting with weights determined via marginalizing over the noise scale. MAGSAC++ is fast -- often an order of magnitude faster -- and more geometrically accurate than MAGSAC. P-NAPSAC merges the advantages of local and global sampling by drawing samples from gradually growing neighborhoods. Exploiting that nearby points are more likely to originate from the same geometric model, P-NAPSAC finds local structures earlier than global samplers. We show that the progressive spatial sampling in P-NAPSAC can be integrated with PROSAC sampling, which is applied to the first, location-defining, point. The methods are tested on homography and fundamental matrix fitting on six publicly available datasets. MAGSAC combined with P-NAPSAC sampler is superior to state-of-the-art robust estimators in terms of speed, accuracy and failure rate.

In most real-world scenarios, labeled training datasets are highly class-imbalanced, where deep neural networks suffer from generalizing to a balanced testing criterion. In this paper, we explore a novel yet simple way to alleviate this issue by augmenting less-frequent classes via translating samples (e.g., images) from more-frequent classes. This simple approach enables a classifier to learn more generalizable features of minority classes, by transferring and leveraging the diversity of the majority information. Our experimental results on a variety of class-imbalanced datasets show that the proposed method improves the generalization on minority classes significantly compared to other existing re-sampling or re-weighting methods. The performance of our method even surpasses those of previous state-of-the-art methods for the imbalanced classification.