CVPR 2019 Papers

Supervised learning methods to infer (hypothesize) depth of a scene from a single image require costly per-pixel ground-truth. We follow a geometric approach that exploits abundant stereo imagery to learn a model to hypothesize scene structure without direct supervision. Although we train a network with stereo pairs, we only require a single image at test time to hypothesize disparity or depth. We propose a novel objective function that exploits the bilateral cyclic relationship between the left and right disparities and we introduce an adaptive regularization scheme that allows the network to handle both the co-visible and occluded regions in a stereo pair. This process ultimately produces a model to generate hypotheses for the 3-dimensional structure of the scene as viewed in a single image. When used to generate a single (most probable) estimate of depth, our method outperforms state-of-the-art unsupervised monocular depth prediction methods on the KITTI benchmarks. We show that our method generalizes well by applying our models trained on KITTI to the Make3d dataset.

Domain adaptation for semantic image segmentation is very necessary since manually labeling large datasets with pixel-level labels is expensive and time consuming. Existing domain adaptation techniques either work on limited datasets, or yield not so good performance compared with supervised learning. In this paper, we propose a novel bidirectional learning framework for domain adaptation of segmentation. Using the bidirectional learning, the image translation model and the segmentation adaptation model can be learned alternatively and promote to each other.Furthermore, we propose a self-supervised learning algorithm to learn a better segmentation adaptation model and in return improve the image translation model. Experiments show that our method superior to the state-of-the-art methods in domain adaptation of segmentation with a big margin. The source code is available at https://github.com/liyunsheng13/BDL

Exploiting multi-scale representations is critical to improve edge detection for objects at different scales. To extract edges at dramatically different scales, we propose a Bi-Directional Cascade Network (BDCN) structure, where an individual layer is supervised by labeled edges at its specific scale, rather than directly applying the same supervision to all CNN outputs. Furthermore, to enrich multi-scale representations learned by BDCN, we introduce a Scale Enhancement Module (SEM) which utilizes dilated convolution to generate multi-scale features, instead of using deeper CNNs or explicitly fusing multi-scale edge maps. These new approaches encourage the learning of multi-scale representations in different layers and detect edges that are well delineated by their scales. Learning scale dedicated layers also results in compact network with a fraction of parameters. We evaluate our method on three datasets, i.e., BSDS500, NYUDv2, and Multicue, and achieve ODS Fmeasure of 0.828, 1.3% higher than current state-of-the art on BSDS500.

Non-line-of-sight (NLOS) imaging is the problem of reconstructing properties of scenes occluded from a sensor, using measurements of light that indirectly travels from the occluded scene to the sensor through intermediate diffuse reflections. We introduce an analysis-by-synthesis framework that can reconstruct complex shape and reflectance of an NLOS object. Our framework deviates from prior work on NLOS reconstruction, by directly optimizing for a surface representation of the NLOS object, in place of commonly employed volumetric representations. At the core of our framework is a new rendering formulation that efficiently computes derivatives of radiometric measurements with respect to NLOS geometry and reflectance, while accurately modeling the underlying light transport physics. By coupling this with stochastic optimization and geometry processing techniques, we are able to reconstruct NLOS surface at a level of detail significantly exceeding what is possible with previous volumetric reconstruction methods.

The simplicity of gradient descent (GD) made it the default method for training ever-deeper and complex neural networks. Both loss functions and architectures are often explicitly tuned to be amenable to this basic local optimization. In the context of weakly-supervised CNN segmentation, we demonstrate a well-motivated loss function where an alternative optimizer (ADM) achieves the state-of-the-art while GD performs poorly. Interestingly, GD obtains its best result for a "smoother" tuning of the loss function. The results are consistent across different network architectures. Our loss is motivated by well-understood MRF/CRF regularization models in "shallow" segmentation and their known global solvers. Our work suggests that network design/training should pay more attention to optimization methods.

As makeup has been widely-adopted for beautification, finding suitable makeup by virtual makeup applications becomes popular. Therefore, a recent line of studies proposes to transfer the makeup from a given reference makeup image to the source non-makeup one. However, it is still challenging due to the massive number of makeup combinations. To facilitate on-demand makeup transfer, in this work, we propose BeautyGlow that decompose the latent vectors of face images derived from the Glow model into makeup and non-makeup latent vectors. Since there is no paired dataset, we formulate a new loss function to guide the decomposition. Afterward, the non-makeup latent vector of a source image and makeup latent vector of a reference image and are effectively combined and revert back to the image domain to derive the results. Experimental results show that the transfer quality of BeautyGlow is comparable to the state-of-the-art methods, while the unique ability to manipulate latent vectors allows BeautyGlow to realize on-demand makeup transfer.

Human action recognition remains as a challenging task partially due to the presence of large variations in the execution of action. To address this issue, we propose a probabilistic model called Hierarchical Dynamic Model (HDM). Leveraging on Bayesian framework, the model parameters are allowed to vary across different sequences of data, which increase the capacity of the model to adapt to intra-class variations on both spatial and temporal extent of actions. Meanwhile, the generative learning process allows the model to preserve the distinctive dynamic pattern for each action class. Through Bayesian inference, we are able to quantify the uncertainty of the classification, providing insight during the decision process. Compared to state-of-the-art methods, our method not only achieves competitive recognition performance within individual dataset but also shows better generalization capability across different datasets. Experiments conducted on data with missing values also show the robustness of the proposed method.

Much of the recent progress made in image classification research can be credited to training procedure refinements, such as changes in data augmentations and optimization methods. In the literature, however, most refinements are either briefly mentioned as implementation details or only visible in source code. In this paper, we will examine a collection of such refinements and empirically evaluate their impact on the final model accuracy through ablation study. We will show that, by combining these refinements together, we are able to improve various CNN models significantly. For example, we raise ResNet-50's top-1 validation accuracy from 75.3% to 79.29% on ImageNet. We will also demonstrate that improvement on image classification accuracy leads to better transfer learning performance in other application domains such as object detection and semantic segmentation.

Deep Convolutional Neural Networks have been adopted for salient object detection and achieved the state-of-the-art performance. Most of the previous works however focus on region accuracy but not on the boundary quality. In this paper, we propose a predict-refine architecture, BASNet, and a new hybrid loss for Boundary-Aware Salient object detection. Specifically, the architecture is composed of a densely supervised Encoder-Decoder network and a residual refinement module, which are respectively in charge of saliency prediction and saliency map refinement. The hybrid loss guides the network to learn the transformation between the input image and the ground truth in a three-level hierarchy -- pixel-, patch- and map- level -- by fusing Binary Cross Entropy (BCE), Structural SIMilarity (SSIM) and Intersection-over-Union (IoU) losses. Equipped with the hybrid loss, the proposed predict-refine architecture is able to effectively segment the salient object regions and accurately predict the fine structures with clear boundaries. Experimental results on six public datasets show that our method outperforms the state-of-the-art methods both in terms of regional and boundary evaluation measures. Our method runs at over 25 fps on a single GPU. The code is available at: https://github.com/NathanUA/BASNet.

This paper presents a novel approach for synthesizing automatically age-progressed facial images in video sequences using Deep Reinforcement Learning. The proposed method models facial structures and the longitudinal face-aging process of given subjects coherently across video frames. The approach is optimized using a long-term reward, Reinforcement Learning function with deep feature extraction from Deep Convolutional Neural Network. Unlike previous age-progression methods that are only able to synthesize an aged likeness of a face from a single input image, the proposed approach is capable of age-progressing facial likenesses in videos with consistently synthesized facial features across frames. In addition, the deep reinforcement learning method guarantees preservation of the visual identity of input faces after age-progression. Results on videos of our new collected aging face AGFW-v2 database demonstrate the advantages of the proposed solution in terms of both quality of age-progressed faces, temporal smoothness, and cross-age face verification.

This work addresses the unsupervised adaptation of an existing object detector to a new target domain. We assume that a large number of unlabeled videos from this domain are readily available. We automatically obtain labels on the target data by using high-confidence detections from the existing detector, augmented with hard (misclassified) examples acquired by exploiting temporal cues using a tracker. These automatically-obtained labels are then used for re-training the original model. A modified knowledge distillation loss is proposed, and we investigate several ways of assigning soft-labels to the training examples from the target domain. Our approach is empirically evaluated on challenging face and pedestrian detection tasks: a face detector trained on WIDER-Face, which consists of high-quality images crawled from the web, is adapted to a large-scale surveillance data set; a pedestrian detector trained on clear, daytime images from the BDD-100K driving data set is adapted to all other scenarios such as rainy, foggy, night-time. Our results demonstrate the usefulness of incorporating hard examples obtained from tracking, the advantage of using soft-labels via distillation loss versus hard-labels, and show promising performance as a simple method for unsupervised domain adaptation of object detectors, with minimal dependence on hyper-parameters.

We introduce the task of scene-aware dialog. Our goal is to generate a complete and natural response to a question about a scene, given video and audio of the scene and the history of previous turns in the dialog. To answer successfully, agents must ground concepts from the question in the video while leveraging contextual cues from the dialog history. To benchmark this task, we introduce the Audio Visual Scene-Aware Dialog (AVSD) Dataset. For each of more than 11,000 videos of human actions from the Charades dataset, our dataset contains a dialog about the video, plus a final summary of the video by one of the dialog participants. We train several baseline systems for this task and evaluate the performance of the trained models using both qualitative and quantitative metrics. Our results indicate that models must utilize all the available inputs (video, audio, question, and dialog history) to perform best on this dataset.

Video-based person re-identification plays an important role in surveillance video analysis, expanding image-based methods by learning features of multiple frames. Most existing methods fuse features by temporal average-pooling, without exploring the different frame weights caused by various viewpoints, poses, and occlusions. In this paper, we propose an attribute-driven method for feature disentangling and frame re-weighting. The features of single frames are disentangled into groups of sub-features, each corresponds to specific semantic attributes. The sub-features are re-weighted by the confidence of attribute recognition and then aggregated at the temporal dimension as the final representation. By means of this strategy, the most informative regions of each frame are enhanced and contributes to a more discriminative sequence representation. Extensive ablation studies demonstrate the effectiveness of feature disentangling as well as temporal re-weighting. The experimental results on the iLIDS-VID, PRID-2011 and MARS datasets demonstrate that our proposed method outperforms existing state-of-the-art approaches.

Since it is difficult to collect face images of the same subject over a long range of age span, most existing face aging methods resort to unpaired datasets to learn age mappings. However, the matching ambiguity between young and aged face images inherent to unpaired training data may lead to unnatural changes of facial attributes during the aging process, which could not be solved by only enforcing identity consistency like most existing studies do. In this paper, we propose an attribute-aware face aging model with wavelet based Generative Adversarial Networks (GANs) to address the above issues. To be specific, we embed facial attribute vectors into both the generator and discriminator of the model to encourage each synthesized elderly face image to be faithful to the attribute of its corresponding input. In addition, a wavelet packet transform (WPT) module is incorporated to improve the visual fidelity of generated images by capturing age-related texture details at multiple scales in the frequency space. Qualitative results demonstrate the ability of our model in synthesizing visually plausible face images, and extensive quantitative evaluation results show that the proposed method achieves state-of-the-art performance on existing datasets.

In this work we address task interference in universal networks by considering that a network is trained on multiple tasks, but performs one task at a time, an approach we refer to as "single-tasking multiple tasks". The network thus modifies its behaviour through task-dependent feature adaptation, or task attention. This gives the network the ability to accentuate the features that are adapted to a task, while shunning irrelevant ones. We further reduce task interference by forcing the task gradients to be statistically indistinguishable through adversarial training, ensuring that the common backbone architecture serving all tasks is not dominated by any of the task-specific gradients. Results in three multi-task dense labelling problems consistently show: (i) a large reduction in the number of parameters while preserving, or even improving performance and (ii) a smooth trade-off between computation and multi-task accuracy. We provide our system's code and pre-trained models at https://github.com/facebookresearch/astmt.

Scene graph generation refers to the task of automatically mapping an image into a semantic structural graph, which requires correctly labeling each extracted object and their interaction relationships. Despite the recent success in object detection using deep learning techniques, inferring complex contextual relationships and structured graph representations from visual data remains a challenging topic. In this study, we propose a novel Attentive Relational Network that consists of two key modules with an object detection backbone to approach this problem. The first module is a semantic transformation module utilized to capture semantic embedded relation features, by translating visual features and linguistic features into a common semantic space. The other module is a graph self-attention module introduced to embed a joint graph representation through assigning various importance weights to neighboring nodes. Finally, accurate scene graphs are produced by the relation inference module to recognize all entities and corresponding relations. We evaluate our proposed method on the widely-adopted Visual Genome Dataset, and the results demonstrate the effectiveness and superiority of our model.

Zero-shot learning (ZSL) aims to classify images from unseen categories, by merely utilizing seen class images as the training data. Existing works on ZSL mainly leverage the global features or learn the global regions, from which, to construct the embeddings to the semantic space. However, few of them study the discrimination power implied in local image regions (parts), which, in some sense, correspond to semantic attributes, have stronger discrimination than attributes, and can thus assist the semantic transfer between seen/unseen classes. In this paper, to discover (semantic) regions, we propose the attentive region embedding network (AREN), which is tailored to advance the ZSL task. Specifically, AREN is end-to-end trainable and consists of two network branches, i.e., the attentive region embedding (ARE) stream, and the attentive compressed second-order embedding (ACSE) stream. ARE is capable of discovering multiple part regions under the guidance of the attention and the compatibility loss. Moreover, a novel adaptive thresholding mechanism is proposed for suppressing redundant (such as background) attention regions. To further guarantee more stable semantic transfer from the perspective of second-order collaboration, ACSE is incorporated into the AREN. In the comprehensive evaluations on four benchmarks, our models achieve state-of-the-art performances under ZSL setting, and compelling results under generalized ZSL setting.

Recent deep learning based salient object detection methods achieve gratifying performance built upon Fully Convolutional Neural Networks (FCNs). However, most of them have suffered from the boundary challenge. The state-of-the-art methods employ feature aggregation tech- nique and can precisely find out wherein the salient object, but they often fail to segment out the entire object with fine boundaries, especially those raised narrow stripes. So there is still a large room for improvement over the FCN based models. In this paper, we design the Attentive Feedback Modules (AFMs) to better explore the structure of objects. A Boundary-Enhanced Loss (BEL) is further employed for learning exquisite boundaries. Our proposed deep model produces satisfying results on the object boundaries and achieves state-of-the-art performance on five widely tested salient object detection benchmarks. The network is in a fully convolutional fashion running at a speed of 26 FPS and does not need any post-processing.

This paper studies panoptic segmentation, a recently proposed task which segments foreground (FG) objects at the instance level as well as background (BG) contents at the semantic level. Existing methods mostly dealt with these two problems separately, but in this paper, we reveal the underlying relationship between them, in particular, FG objects provide complementary cues to assist BG understanding. Our approach, named the Attention-guided Unified Network (AUNet), is a unified framework with two branches for FG and BG segmentation simultaneously. Two sources of attentions are added to the BG branch, namely, RPN and FG segmentation mask to provide object-level and pixel-level attentions, respectively. Our approach is generalized to different backbones with consistent accuracy gain in both FG and BG segmentation, and also sets new state-of-the-arts both in the MS-COCO (46.5% PQ) and Cityscapes (59.0% PQ) benchmarks.