CVPR 2020 论文列表

We consider the problem of optimizing the performance of an active imaging system by automatically discovering the illuminations it should use, and the way to decode them. Our approach tackles two seemingly incompatible goals: (1) "tuning" the illuminations and decoding algorithm precisely to the devices at hand---to their optical transfer functions, non-linearities, spectral responses, image processing pipelines---and (2) doing so without modeling or calibrating the system; without modeling the scenes of interest; and without prior training data. The key idea is to formulate a stochastic gradient descent (SGD) optimization procedure that puts the actual system in the loop: projecting patterns, capturing images, and calculating the gradient of expected reconstruction error. We apply this idea to structured-light triangulation to "auto-tune" several devices---from smartphones and laser projectors to advanced computational cameras. Our experiments show that despite being model-free and automatic, optical SGD can boost system 3D accuracy substantially over state-of-the-art coding schemes.

Conventional unsupervised hashing methods usually take advantage of similarity graphs, which are either pre-computed in the high-dimensional space or obtained from random anchor points. On the one hand, existing methods uncouple the procedures of hash function learning and graph construction. On the other hand, graphs empirically built upon original data could introduce biased prior knowledge of data relevance, leading to sub-optimal retrieval performance. In this paper, we tackle the above problems by proposing an efficient and adaptive code-driven graph, which is updated by decoding in the context of an auto-encoder. Specifically, we introduce into our framework twin bottlenecks (i.e., latent variables) that exchange crucial information collaboratively. One bottleneck (i.e., binary codes) conveys the high-level intrinsic data structure captured by the code-driven graph to the other (i.e., continuous variables for low-level detail information), which in turn propagates the updated network feedback for the encoder to learn more discriminative binary codes. The auto-encoding learning objective literally rewards the code-driven graph to learn an optimal encoder. Moreover, the proposed model can be simply optimized by gradient descent without violating the binary constraints. Experiments on benchmarked datasets clearly show the superiority of our framework over the state-of-the-art hashing methods. Our source code can be found at https://github.com/ymcidence/TBH.

Colorectal cancer screening modalities, such as optical colonoscopy (OC) and virtual colonoscopy (VC), are critical for diagnosing and ultimately removing polyps (precursors for colon cancer). The non-invasive VC is normally used to inspect a 3D reconstructed colon (from computed tomography scans) for polyps and if found, the OC procedure is performed to physically traverse the colon via endoscope and remove these polyps. In this paper, we present a deep learning framework, Extended and Directional CycleGAN, for lossy unpaired image-to-image translation between OC and VC to augment OC video sequences with scale-consistent depth information from VC and VC with patient-specific textures, color and specular highlights from OC (e.g. for realistic polyp synthesis). Both OC and VC contain structural information, but it is obscured in OC by additional patient-specific texture and specular highlights, hence making the translation from OC to VC lossy. The existing CycleGAN approaches do not handle lossy transformations. To address this shortcoming, we introduce an extended cycle consistency loss, which compares the geometric structures from OC in the VC domain. This loss removes the need for the CycleGAN to embed OC information in the VC domain. To handle a stronger removal of the textures and lighting, a Directional Discriminator is introduced to differentiate the direction of translation (by creating paired information for the discriminator), as opposed to the standard CycleGAN which is direction-agnostic. Combining the extended cycle consistency loss and the Directional Discriminator, we show state-of-the-art results on scale-consistent depth inference for phantom, textured VC and for real polyp and normal colon video sequences. We also present results for realistic pendunculated and flat polyp synthesis from bumps introduced in 3D VC models.

Large-batch SGD is important for scaling training of deep neural networks. However, without fine-tuning hyperparameter schedules, the generalization of the model may be hampered. We propose to use batch augmentation: replicating instances of samples within the same batch with different data augmentations. Batch augmentation acts as a regularizer and an accelerator, increasing both generalization and performance scaling for a fixed budget of optimization steps. We analyze the effect of batch augmentation on gradient variance and show that it empirically improves convergence for a wide variety of networks and datasets. Our results show that batch augmentation reduces the number of necessary SGD updates to achieve the same accuracy as the state-of-the-art. Overall, this simple yet effective method enables faster training and better generalization by allowing more computational resources to be used concurrently.

Current state-of-the-art detectors typically exploit feature pyramid to detect objects at different scales. Among them, FPN is one of the representative works that build a feature pyramid by multi-scale features summation. However, the design defects behind prevent the multi-scale features from being fully exploited. In this paper, we begin by first analyzing the design defects of feature pyramid in FPN, and then introduce a new feature pyramid architecture named AugFPN to address these problems. Specifically, AugFPN consists of three components: Consistent Supervision, Residual Feature Augmentation, and Soft RoI Selection. AugFPN narrows the semantic gaps between features of different scales before feature fusion through Consistent Supervision. In feature fusion, ratio-invariant context information is extracted by Residual Feature Augmentation to reduce the information loss of feature map at the highest pyramid level. Finally, Soft RoI Selection is employed to learn a better RoI feature adaptively after feature fusion. By replacing FPN with AugFPN in Faster R-CNN, our models achieve 2.3 and 1.6 points higher Average Precision (AP) when using ResNet50 and MobileNet-v2 as backbone respectively. Furthermore, AugFPN improves RetinaNet by 1.6 points AP and FCOS by 0.9 points AP when using ResNet50 as backbone. Codes are available on https://github.com/Gus-Guo/AugFPN.

We study the attribution problem for deep networks applied to perception tasks. For vision tasks, attribution techniques attribute the prediction of a network to the pixels of the input image. We propose a new technique called Blur Integrated Gradients (Blur IG). This technique has several advantages over other methods. First, it can tell at what scale a network recognizes an object. It produces scores in the scale/frequency dimension, that we find captures interesting phenomena. Second, it satisfies the scale-space axioms, which imply that it employs perturbations that are free of artifact. We therefore produce explanations that are cleaner and consistent with the operation of deep networks. Third, it eliminates the need for baseline parameter for Integrated Gradients for perception tasks. This is desirable because the choice of baseline has a significant effect on the explanations. We compare the proposed technique against previous techniques and demonstrate application on three tasks: ImageNet object recognition, Diabetic Retinopathy prediction, and AudioSet audio event identification. Code and examples are at https://github.com/PAIR-code/saliency.

Few shot image classification aims at learning a classifier from limited labeled data. Generating the classification weights has been applied in many meta-learning methods for few shot image classification due to its simplicity and effectiveness. In this work, we present Attentive Weights Generation for few shot learning via Information Maximization (AWGIM), which introduces two novel contributions: i) Mutual information maximization between generated weights and data within the task; this enables the generated weights to retain information of the task and the specific query sample. ii) Self-attention and cross-attention paths to encode the context of the task and individual queries. Both two contributions are shown to be very effective in extensive experiments. Overall, AWGIM is competitive with state-of-the-art. Code is available at https://github.com/Yiluan/AWGIM.

Traditional convolution-based generative adversarial networks synthesize images based on hierarchical local operations, where long-range dependency relation is implicitly modeled with a Markov chain. It is still not sufficient for categories with complicated structures. In this paper, we characterize long-range dependence with attentive normalization (AN), which is an extension to traditional instance normalization. Specifically, the input feature map is softly divided into several regions based on its internal semantic similarity, which are respectively normalized. It enhances consistency between distant regions with semantic correspondence. Compared with self-attention GAN, our attentive normalization does not need to measure the correlation of all locations, and thus can be directly applied to large-size feature maps without much computational burden. Extensive experiments on class-conditional image generation and semantic inpainting verify the efficacy of our proposed module.

Existing deep learning based matting algorithms primarily resort to high-level semantic features to improve the overall structure of alpha mattes. However, we argue that advanced semantics extracted from CNNs contribute unequally for alpha perception and we are supposed to reconcile advanced semantic information with low-level appearance cues to refine the foreground details. In this paper, we propose an end-to-end Hierarchical Attention Matting Network (HAttMatting), which can predict the better structure of alpha mattes from single RGB images without additional input. Specifically, we employ spatial and channel-wise attention to integrate appearance cues and pyramidal features in a novel fashion. This blended attention mechanism can perceive alpha mattes from refined boundaries and adaptive semantics. We also introduce a hybrid loss function fusing Structural SIMilarity (SSIM), Mean Square Error (MSE) and Adversarial loss to guide the network to further improve the overall foreground structure. Besides, we construct a large-scale image matting dataset comprised of 59,600 training images and 1000 test images (total 646 distinct foreground alpha mattes), which can further improve the robustness of our hierarchical structure aggregation model. Extensive experiments demonstrate that the proposed HAttMatting can capture sophisticated foreground structure and achieve state-of-the-art performance with single RGB images as input.

Facial landmark detection is a fundamental task for many consumer and high-end applications and is almost entirely solved by machine learning methods today. Existing datasets used to train such algorithms are primarily made up of only low resolution images, and current algorithms are limited to inputs of comparable quality and resolution as the training dataset. On the other hand, high resolution imagery is becoming increasingly more common as consumer cameras improve in quality every year. Therefore, there is need for algorithms that can leverage the rich information available in high resolution imagery. Naively attempting to reuse existing network architectures on high resolution imagery is prohibitive due to memory bottlenecks on GPUs. The only current solution is to downsample the images, sacrificing resolution and quality. Building on top of recent progress in attention-based networks, we present a novel, fully convolutional regional architecture that is specially designed for predicting landmarks on very high resolution facial images without downsampling. We demonstrate the flexibility of our architecture by training the proposed model with images of resolutions ranging from 256 x 256 to 4K. In addition to being the first method for facial landmark detection on high resolution images, our approach achieves superior performance over traditional (holistic) state-of-the-art architectures across ALL resolutions, leading to a general-purpose, extremely flexible, high quality landmark detector.

Situation Recognition (SR) is a fine-grained action recognition task where the model is expected to not only predict the salient action of the image, but also predict values of all associated semantic roles of the action. Predicting semantic roles is very challenging: a vast variety of possibilities can be the match for a semantic role. Existing work has focused on dependency modelling architectures to solve this issue. Inspired by the success achieved by query-based visual reasoning (e.g., Visual Question Answering), we propose to address semantic role prediction as a query-based visual reasoning problem. However, existing query-based reasoning methods have not considered handling of inter-dependent queries which is a unique requirement of semantic role prediction in SR. Therefore, to the best of our knowledge, we propose the first set of methods to address inter-dependent queries in query-based visual reasoning. Extensive experiments demonstrate the effectiveness of our proposed method which achieves outstanding performance on Situation Recognition task. Furthermore, leveraging query inter-dependency, our methods improve upon a state-of-the-art method that answers queries separately. Our code: https://github.com/thilinicooray/context-aware-reasoning-for-sr

Multi-view stereo is a crucial task in computer vision, that requires accurate and robust photo-consistency among input images for depth estimation. Recent studies have shown that learning-based feature matching and confidence regularization can play a vital role in this task. Nevertheless, how to design good matching confidence volumes as well as effective regularizers for them are still under in-depth study. In this paper, we propose an attention-aware deep neural network "AttMVS" for learning multi-view stereo. In particular, we propose a novel attention-enhanced matching confidence volume, that combines the raw pixel-wise matching confidence from the extracted perceptual features with the contextual information of local scenes, to improve the matching robustness. Furthermore, we develop an attention-guided regularization module, which consists of multilevel ray fusion modules, to hierarchically aggregate and regularize the matching confidence volume into a latent depth probability volume.Experimental results show that our approach achieves the best overall performance on the DTU dataset and the intermediate sequences of Tanks & Temples benchmark over many state-of-the-art MVS algorithms.

Convolutional Neural Network (CNN) based methods generally take crowd counting as a regression task by outputting crowd densities. They learn the mapping between image contents and crowd density distributions. Though having achieved promising results, these data-driven counting networks are prone to overestimate or underestimate people counts of regions with different density patterns, which degrades the whole count accuracy. To overcome this problem, we propose an approach to alleviate the counting performance differences in different regions. Specifically, our approach consists of two networks named Density Attention Network (DANet) and Attention Scaling Network (ASNet). DANet provides ASNet with attention masks related to regions of different density levels. ASNet first generates density maps and scaling factors and then multiplies them by attention masks to output separate attention-based density maps. These density maps are summed to give the final density map. The attention scaling factors help attenuate the estimation errors in different regions. Furthermore, we present a novel Adaptive Pyramid Loss (APLoss) to hierarchically calculate the estimation losses of sub-regions, which alleviates the training bias. Extensive experiments on four challenging datasets (ShanghaiTech Part A, UCF_CC_50, UCF-QNRF, and WorldExpo'10) demonstrate the superiority of the proposed approach.

We propose a novel attention-based framework for 3D human pose estimation from a monocular video. Despite the general success of end-to-end deep learning paradigms, our approach is based on two key observations: (1) temporal incoherence and jitter are often yielded from a single frame prediction; (2) error rate can be remarkably reduced by increasing the receptive field in a video. Therefore, we design an attentional mechanism to adaptively identify significant frames and tensor outputs from each deep neural net layer, leading to a more optimal estimation. To achieve large temporal receptive fields, multi-scale dilated convolutions are employed to model long-range dependencies among frames. The architecture is straightforward to implement and can be flexibly adopted for real-time applications. Any off-the-shelf 2D pose estimation system, e.g. Mocap libraries, can be easily integrated in an ad-hoc fashion. We both quantitatively and qualitatively evaluate our method on various standard benchmark datasets (e.g. Human3.6M, HumanEva). Our method considerably outperforms all the state-of-the-art algorithms up to 8% error reduction (average mean per joint position error: 34.7) as compared to the best-reported results. Code is available at: (https://github.com/lrxjason/Attention3DHumanPose)

Fine-grained visual categorization (FGVC) is an important but challenging task due to high intra-class variances and low inter-class variances caused by deformation, occlusion, illumination, etc. An attention convolutional binary neural tree architecture is presented to address those problems for weakly supervised FGVC. Specifically, we incorporate convolutional operations along edges of the tree structure, and use the routing functions in each node to determine the root-to-leaf computational paths within the tree. The final decision is computed as the summation of the predictions from leaf nodes. The deep convolutional operations learn to capture the representations of objects, and the tree structure characterizes the coarse-to-fine hierarchical feature learning process. In addition, we use the attention transformer module to enforce the network to capture discriminative features. The negative log-likelihood loss is used to train the entire network in an end-to-end fashion by SGD with back-propagation. Several experiments on the CUB-200-2011, Stanford Cars and Aircraft datasets demonstrate that the proposed method performs favorably against the state-of-the-arts.

Deep visual models are susceptible to extremely low magnitude perturbations to input images. Though carefully crafted, the perturbation patterns generally appear noisy, yet they are able to perform controlled manipulation of model predictions. This observation is used to argue that deep representation is misaligned with human perception. This paper counter-argues and proposes the first attack on deep learning that aims at explaining the learned representation instead of fooling it. By extending the input domain of the manipulative signal and employing a model faithful channelling, we iteratively accumulate adversarial perturbations for a deep model. The accumulated signal gradually manifests itself as a collection of visually salient features of the target label (in model fooling), casting adversarial perturbations as primitive features of the target label. Our attack provides the first demonstration of systematically computing perturbations for adversarially non-robust classifiers that comprise salient visual features of objects. We leverage the model explaining character of our algorithm to perform image generation, inpainting and interactive image manipulation by attacking adversarially robust classifiers. The visually appealing results across these applications demonstrate the utility of our attack (and perturbations in general) beyond model fooling.

Object detection from 3D point clouds remains a challenging task, though recent studies pushed the envelope with the deep learning techniques. Owing to the severe spatial occlusion and inherent variance of point density with the distance to sensors, appearance of a same object varies a lot in point cloud data. Designing robust feature representation against such appearance changes is hence the key issue in a 3D object detection method. In this paper, we innovatively propose a domain adaptation like approach to enhance the robustness of the feature representation. More specifically, we bridge the gap between the perceptual domain where the feature comes from a real scene and the conceptual domain where the feature is extracted from an augmented scene consisting of non-occlusion point cloud rich of detailed information. This domain adaptation approach mimics the functionality of the human brain when proceeding object perception. Extensive experiments demonstrate that our simple yet effective approach fundamentally boosts the performance of 3D point cloud object detection and achieves the state-of-the-art results.

We introduce a new large-scale dataset that links the assessment of image quality issues to two practical vision tasks: image captioning and visual question answering. First, we identify for 39,181 images taken by people who are blind whether each is sufficient quality to recognize the content as well as what quality flaws are observed from six options. These labels serve as a critical foundation for us to make the following contributions: (1) a new problem and algorithms for deciding whether an image is insufficient quality to recognize the content and so not captionable, (2) a new problem and algorithms for deciding which of six quality flaws an image contains, (3) a new problem and algorithms for deciding whether a visual question is unanswerable due to unrecognizable content versus the content of interest being missing from the field of view, and (4) a novel application of more efficiently creating a large-scale image captioning dataset by automatically deciding whether an image is insufficient quality and so should not be captioned. We publicly-share our datasets and code to facilitate future extensions of this work: https://vizwiz.org.

With the wide use of artistic images, aesthetic quality assessment has been widely concerned. How to integrate aesthetics into image editing is still a problem worthy of discussion. In this paper, aesthetic assessment is introduced into eye in-painting task for the first time. We construct an eye aesthetic dataset, and train the eye aesthetic assessment network on this basis. Then we propose a novel eye aesthetic and face semantic guided multi-reference eye inpainting GAN approach (AesGAN), which automatically selects the best reference under the guidance of eye aesthetics. A new aesthetic loss has also been introduced into the network to learn the eye aesthetic features and generate highquality eyes. We prove the effectiveness of eye aesthetic assessment in our experiments, which may inspire more applications of aesthetics assessment. Both qualitative and quantitative experimental results show that the proposed AesGAN can produce more natural and visually attractive eyes compared with state-of-the-art methods.