CVPR 2020 论文列表

Convolutional neural networks (CNNs) achieve translational invariance by using pooling operations. However, the operations do not preserve the spatial relationships in the learned representations. Hence, CNNs cannot extrapolate to various geometric transformations of inputs. Recently, Capsule Networks (CapsNets) have been proposed to tackle this problem. In CapsNets, each entity is represented by a vector and routed to high-level entity representations by a dynamic routing algorithm. CapsNets have been shown to be more robust than CNNs to affine transformations of inputs. However, there is still a huge gap between their performance on transformed inputs compared to untransformed versions. In this work, we first revisit the routing procedure by (un)rolling its forward and backward passes. Our investigation reveals that the routing procedure contributes neither to the generalization ability nor to the affine robustness of the CapsNets. Furthermore, we explore the limitations of capsule transformations and propose affine CapsNets (Aff-CapsNets), which are more robust to affine transformations. On our benchmark task, where models are trained on the MNIST dataset and tested on the AffNIST dataset, our Aff-CapsNets improve the benchmark performance by a large margin (from 79% to 93.21%), without using any routing mechanism.

Neural architecture search (NAS) has demonstrated much success in automatically designing effective neural network architectures. To improve the efficiency of NAS, previous approaches adopt weight sharing method to force all models share the same set of weights. However, it has been observed that a model performing better with shared weights does not necessarily perform better when trained alone. In this paper, we analyse existing weight sharing one-shot NAS approaches from a Bayesian point of view and identify the Posterior Fading problem, which compromises the effectiveness of shared weights. To alleviate this problem, we present a novel approach to guide the parameter posterior towards its true distribution. Moreover, a hard latency constraint is introduced during the search so that the desired latency can be achieved. The resulted method, namely Posterior Convergent NAS (PC-NAS), achieves state-of-the-art performance under standard GPU latency constraint on ImageNet.

Recent advances on CNNs are mostly devoted to designing more complex architectures to enhance their representation learning capacity. In this paper, we consider how to improve the basic convolutional feature transformation process of CNNs without tuning the model architectures. To this end, we present a novel self-calibrated convolutions that explicitly expand fields-of-view of each convolutional layers through internal communications and hence enrich the output features. In particular, unlike the standard convolutions that fuse spatial and channel-wise information using small kernels (e.g., 3x3), self-calibrated convolutions adaptively build long-range spatial and inter-channel dependencies around each spatial location through a novel self-calibration operation. Thus, it can help CNNs generate more discriminative representations by explicitly incorporating richer information. Our self-calibrated convolution design is simple and generic, and can be easily applied to augment standard convolutional layers without introducing extra parameters and complexity. Extensive experiments demonstrate that when applying self-calibrated convolutions into different backbones, our networks can significantly improve the baseline models in a variety of vision tasks, including image recognition, object detection, instance segmentation, and keypoint detection, with no need to change the network architectures. We hope this work could provide a promising way for future research in designing novel convolutional feature transformations for improving convolutional networks. Code is available on the project page.

Temporal relations among multiple action segments play an important role in action segmentation especially when observations are limited (e.g., actions are occluded by other objects or happen outside a field of view). In this paper, we propose a network module called Graph-based Temporal Reasoning Module (GTRM) that can be built on top of existing action segmentation models to learn the relation of multiple action segments in various time spans. We model the relations by using two Graph Convolution Networks (GCNs) where each node represents an action segment. The two graphs have different edge properties to account for boundary regression and classification tasks, respectively. By applying graph convolution, we can update each node's representation based on its relation with neighboring nodes. The updated representation is then used for improved action segmentation. We evaluate our model on the challenging egocentric datasets namely EGTEA and EPIC-Kitchens, where actions may be partially observed due to the viewpoint restriction. The results show that our proposed GTRM outperforms state-of-the-art action segmentation models by a large margin. We also demonstrate the effectiveness of our model on two third-person video datasets, the 50Salads dataset and the Breakfast dataset.

Few-shot learning is a fundamental task in computer vision that carries the promise of alleviating the need for exhaustively labeled data. Most few-shot learning approaches to date have focused on progressively more complex neural feature extractors and classifier adaptation strategies, and the refinement of the task definition itself. In this paper, we explore the hypothesis that a simple class-covariance-based distance metric, namely the Mahalanobis distance, adopted into a state of the art few-shot learning approach (CNAPS) can, in and of itself, lead to a significant performance improvement. We also discover that it is possible to learn adaptive feature extractors that allow useful estimation of the high dimensional feature covariances required by this metric from surprisingly few samples. The result of our work is a new "Simple CNAPS" architecture which has up to 9.2% fewer trainable parameters than CNAPS and performs up to 6.1% better than state of the art on the standard few-shot image classification benchmark dataset.

While many works focus on 3D reconstruction from images, in this paper, we focus on 3D shape reconstruction and completion from a variety of 3D inputs, which are deficient in some respect: low and high resolution voxels, sparse and dense point clouds, complete or incomplete. Processing of such 3D inputs is an increasingly important problem as they are the output of 3D scanners, which are becoming more accessible, and are the intermediate output of 3D computer vision algorithms. Recently, learned implicit functions have shown great promise as they produce continuous reconstructions. However, we identified two limitations in reconstruction from 3D inputs: 1) details present in the input data are not retained, and 2) poor reconstruction of articulated humans. To solve this, we propose Implicit Feature Networks (IF-Nets), which deliver continuous outputs, can handle multiple topologies, and complete shapes for missing or sparse input data retaining the nice properties of recent learned implicit functions, but critically they can also retain detail when it is present in the input data, and can reconstruct articulated humans. Our work differs from prior work in two crucial aspects. First, instead of using a single vector to encode a 3D shape, we extract a learnable 3-dimensional multi-scale tensor of deep features, which is aligned with the original Euclidean space embedding the shape. Second, instead of classifying x-y-z point coordinates directly, we classify deep features extracted from the tensor at a continuous query point. We show that this forces our model to make decisions based on global and local shape structure, as opposed to point coordinates, which are arbitrary under Euclidean transformations. Experiments demonstrate that IF-Nets outperform prior work in 3D object reconstruction in ShapeNet, and obtain significantly more accurate 3D human reconstructions. Code and project website is available at https://virtualhumans.mpi-inf.mpg.de/ifnets/.

Trajectory forecasting and imputation are pivotal steps towards understanding the movement of human and objects, which are quite challenging since the future trajectories and missing values in a temporal sequence are full of uncertainties, and the spatial-temporally contextual correlation is hard to model. Yet, the relevance between sequence prediction and imputation is disregarded by existing approaches. To this end, we propose a novel imitative non-autoregressive modeling method to simultaneously handle the trajectory prediction task and the missing value imputation task. Specifically, our framework adopts an imitation learning paradigm, which contains a recurrent conditional variational autoencoder (RC-VAE) as a demonstrator, and a non-autoregressive transformation model (NART) as a learner. By jointly optimizing the two models, RC-VAE can predict the future trajectory and capture the temporal relationship in the sequence to supervise the NART learner. As a result, NART learns from the demonstrator and imputes the missing value in a non autoregressive strategy. We conduct extensive experiments on three popular datasets, and the results show that our model achieves state-of-the-art performance across all the datasets.

We propose Image2StyleGAN++, a flexible image editing framework with many applications. Our framework extends the recent Image2StyleGAN in three ways. First, we introduce noise optimization as a complement to the W+ latent space embedding. Our noise optimization can restore high frequency features in images and thus significantly improves the quality of reconstructed images, e.g. a big increase of PSNR from 20 dB to 45 dB. Second, we extend the global W+ latent space embedding to enable local embeddings. Third, we combine embedding with activation tensor manipulation to perform high quality local edits along with global semantic edits on images. Such edits motivate various high quality image editing applications, e.g. image reconstruction, image inpainting, image crossover, local style transfer, image editing using scribbles, and attribute level feature transfer. Examples of the edited images are shown across the paper for visual inspection.

Deep convolution-based single image super-resolution (SISR) networks embrace the benefits of learning from large-scale external image resources for local recovery, yet most existing works have ignored the long-range feature-wise similarities in natural images. Some recent works have successfully leveraged this intrinsic feature correlation by exploring non-local attention modules. However, none of the current deep models have studied another inherent property of images: cross-scale feature correlation. In this paper, we propose the first Cross-Scale Non-Local (CS-NL) attention module with integration into a recurrent neural network. By combining the new CS-NL prior with local and in-scale non-local priors in a powerful recurrent fusion cell, we can find more cross-scale feature correlations within a single low-resolution (LR) image. The performance of SISR is significantly improved by exhaustively integrating all possible priors. Extensive experiments demonstrate the effectiveness of the proposed CS-NL module by setting new state-of-the-arts on multiple SISR benchmarks.

Image search with text feedback has promising impacts in various real-world applications, such as e-commerce and internet search. Given a reference image and text feedback from user, the goal is to retrieve images that not only resemble the input image, but also change certain aspects in accordance with the given text. This is a challenging task as it requires the synergistic understanding of both image and text. In this work, we tackle this task by a novel Visiolinguistic Attention Learning (VAL) framework. Specifically, we propose a composite transformer that can be seamlessly plugged in a CNN to selectively preserve and transform the visual features conditioned on language semantics. By inserting multiple composite transformers at varying depths, VAL is incentive to encapsulate the multi-granular visiolinguistic information, thus yielding an expressive representation for effective image search. We conduct comprehensive evaluation on three datasets: Fashion200k, Shoes and FashionIQ. Extensive experiments show our model exceeds existing approaches on all datasets, demonstrating consistent superiority in coping with various text feedbacks, including attribute-like and natural language descriptions.

Despite the success of Generative Adversarial Networks (GANs) in image synthesis, applying trained GAN models to real image processing remains challenging. Previous methods typically invert a target image back to the latent space either by back-propagation or by learning an additional encoder. However, the reconstructions from both of the methods are far from ideal. In this work, we propose a novel approach, called mGANprior, to incorporate the well-trained GANs as effective prior to a variety of image processing tasks. In particular, we employ multiple latent codes to generate multiple feature maps at some intermediate layer of the generator, then compose them with adaptive channel importance to recover the input image. Such an over-parameterization of the latent space significantly improves the image reconstruction quality, outperforming existing competitors. The resulting high-fidelity image reconstruction enables the trained GAN models as prior to many real-world applications, such as image colorization, super-resolution, image inpainting, and semantic manipulation. We further analyze the properties of the layer-wise representation learned by GAN models and shed light on what knowledge each layer is capable of representing.

Image demoireing is a multi-faceted image restoration task involving both texture and color restoration. In this paper, we propose a novel multiscale bandpass convolutional neural network (MBCNN) to address this problem. As an end-to-end solution, MBCNN respectively solves the two sub-problems. For texture restoration, we propose a learnable bandpass filter (LBF) to learn the frequency prior for moire texture removal. For color restoration, we propose a two-step tone mapping strategy, which first applies a global tone mapping to correct for a global color shift, then performs local fine tuning of the color per pixel. Through an ablation study, we demonstrate the effectiveness of the different components of MBCNN. Experimental results on two public datasets show that our method outperforms state-of-the-art methods by a large margin (more than 2dB in terms of PSNR).

3D object detection has seen quick progress thanks to advances in deep learning on point clouds. A few recent works have even shown state-of-the-art performance with just point clouds input (e.g. VoteNet). However, point cloud data have inherent limitations. They are sparse, lack color information and often suffer from sensor noise. Images, on the other hand, have high resolution and rich texture. Thus they can complement the 3D geometry provided by point clouds. Yet how to effectively use image information to assist point cloud based detection is still an open question. In this work, we build on top of VoteNet and propose a 3D detection architecture called ImVoteNet specialized for RGB-D scenes. ImVoteNet is based on fusing 2D votes in images and 3D votes in point clouds. Compared to prior work on multi-modal detection, we explicitly extract both geometric and semantic features from the 2D images. We leverage camera parameters to lift these features to 3D. To improve the synergy of 2D-3D feature fusion, we also propose a multi-tower training scheme. We validate our model on the challenging SUN RGB-D dataset, advancing state-of-the-art results by 5.7 mAP. We also provide rich ablation studies to analyze the contribution of each design choice.

Enabling bi-directional retrieval of images and texts is important for understanding the correspondence between vision and language. Existing methods leverage the attention mechanism to explore such correspondence in a fine-grained manner. However, most of them consider all semantics equally and thus align them uniformly, regardless of their diverse complexities. In fact, semantics are diverse (i.e. involving different kinds of semantic concepts), and humans usually follow a latent structure to combine them into understandable languages. It may be difficult to optimally capture such sophisticated correspondences in existing methods. In this paper, to address such a deficiency, we propose an Iterative Matching with Recurrent Attention Memory (IMRAM) method, in which correspondences between images and texts are captured with multiple steps of alignments. Specifically, we introduce an iterative matching scheme to explore such fine-grained correspondence progressively. A memory distillation unit is used to refine alignment knowledge from early steps to later ones. Experiment results on three benchmark datasets, i.e. Flickr8K, Flickr30K, and MS COCO, show that our IMRAM achieves state-of-the-art performance, well demonstrating its effectiveness. Experiments on a practical business advertisement dataset, named KWAI-AD, further validates the applicability of our method in practical scenarios.

With the increasing number of layers and parameters in neural networks, the energy consumption of neural networks has become a great concern to society, especially to users of handheld or embedded devices. In this paper, we investigate the robustness of neural networks against energy-oriented attacks. Specifically, we propose ILFO (Intermediate Output-Based Loss Function Optimization) attack against a common type of energy-saving neural networks, Adaptive Neural Networks (AdNN). AdNNs save energy consumption by dynamically deactivating part of its model based on the need of the inputs. ILFO leverages intermediate output as a proxy to infer the relation between input and its corresponding energy consumption. ILFO has shown an increase up to 100 % of the FLOPs (floating-point operations per second) reduced by AdNNs with minimum noise added to input images. To our knowledge, this is the first attempt to attack the energy consumption of an AdNN.

3D object detection is an important scene understanding task in autonomous driving and virtual reality. Approaches based on LiDAR technology have high performance, but LiDAR is expensive. Considering more general scenes, where there is no LiDAR data in the 3D datasets, we propose a 3D object detection approach from stereo vision which does not rely on LiDAR data either as input or as supervision in training, but solely takes RGB images with corresponding annotated 3D bounding boxes as training data. As depth estimation of object is the key factor affecting the performance of 3D object detection, we introduce an Instance-DepthAware (IDA) module which accurately predicts the depth of the 3D bounding box's center by instance-depth awareness, disparity adaptation and matching cost reweighting. Moreover, our model is an end-to-end learning framework which does not require multiple stages or postprocessing algorithm. We provide detailed experiments on KITTI benchmark and achieve impressive improvements compared with the existing image-based methods. Our code is available at https://github.com/swords123/IDA-3D.

One of the fundamental problems that arise in multimodal learning tasks is the disparity of information levels between different modalities. To resolve this problem, we propose Hypergraph Attention Networks (HANs), which define a common semantic space among the modalities with symbolic graphs and extract a joint representation of the modalities based on a co-attention map constructed in the semantic space. HANs follow the process: constructing the common semantic space with symbolic graphs of each modality, matching the semantics between sub-structures of the symbolic graphs, constructing co-attention maps between the graphs in the semantic space, and integrating the multimodal inputs using the co-attention maps to get the final joint representation. From the qualitative analysis with two Visual Question and Answering datasets, we discover that 1) the alignment of the information levels between the modalities is important, and 2) the symbolic graphs are very powerful ways to represent the information of the low-level signals in alignment. Moreover, HANs dramatically improve the state-of-the-art accuracy on the GQA dataset from 54.6% to 61.88% only using the symbolic information in quantitatively.

This paper proposes a Hyperbolic Visual Embedding Learning Network for zero-shot recognition. The network learns image embeddings in hyperbolic space, which is capable of preserving the hierarchical structure of semantic classes in low dimensions. Comparing with existing zero-shot learning approaches, the network is more robust because the embedding feature in hyperbolic space better represents class hierarchy and thereby avoid misleading resulted from unrelated siblings. Our network outperforms exiting baselines under hierarchical evaluation with an extremely challenging setting, i.e., learning only from 1,000 categories to recognize 20,841 unseen categories. While under flat evaluation, it has competitive performance as state-of-the-art methods but with five times lower embedding dimensions. Our code is publicly available (https://github.com/ShaoTengLiu/Hyperbolic_ZSL).

We introduce HybridPose, a novel 6D object pose estimation approach. HybridPose utilizes a hybrid intermediate representation to express different geometric information in the input image, including keypoints, edge vectors, and symmetry correspondences. Compared to a unitary representation, our hybrid representation allows pose regression to exploit more and diverse features when one type of predicted representation is inaccurate (e.g., because of occlusion). Different intermediate representations used by HybridPose can all be predicted by the same simple neural network, and outliers in predicted intermediate representations are filtered by a robust regression module. Compared to state-of-the-art pose estimation approaches, HybridPose is comparable in running time and is significantly more accurate. For example, on Occlusion Linemod dataset, our method achieves a prediction speed of 30 fps with a mean ADD(-S) accuracy of 79.2%, representing a 67.4% improvement from the current state-of-the-art approach.