CVPR 2019 Papers

Low-rank matrix completion (LRMC) is a classical model in both computer vision (CV) and machine learning, and has been successfully applied to various real applications. In the recent CV tasks, the completion is usually employed on the variants of data, such as "non-local" or filtered, rather than their original forms. This fact makes that the theoretical analysis of the conventional LRMC is no longer suitable in these applications. To tackle this problem, we propose a more general framework for LRMC, in which the linear transformations of the data are taken into account. We rigorously prove the identifiability of the proposed model and show an upper bound of the reconstruction error. Furthermore, we derive an efficient completion algorithm by using augmented Lagrangian multipliers and the sketching trick. In the experiments, we apply the proposed method to the classical image inpainting problem and achieve the state-of-the-art results.

Video description is one of the most challenging problems in vision and language understanding due to the large variability both on the video and language side. Models, hence, typically shortcut the difficulty in recognition and generate plausible sentences that are based on priors but are not necessarily grounded in the video. In this work, we explicitly link the sentence to the evidence in the video by annotating each noun phrase in a sentence with the corresponding bounding box in one of the frames of a video. Our dataset, ActivityNet-Entities, augments the challenging ActivityNet Captions dataset with 158k bounding box annotations, each grounding a noun phrase. This allows training video description models with this data, and importantly, evaluate how grounded or "true" such model are to the video they describe. To generate grounded captions, we propose a novel video description model which is able to exploit these bounding box annotations. We demonstrate the effectiveness of our model on our dataset, but also show how it can be applied to image description on the Flickr30k Entities dataset. We achieve state-of-the-art performance on video description, video paragraph description, and image description and demonstrate our generated sentences are better grounded in the video.

Tracking by siamese networks has achieved favorable performance in recent years. However, most of existing siamese methods do not take full advantage of spatial-temporal target appearance modeling under different contextual situations. In fact, the spatial-temporal information can provide diverse features to enhance the target representation, and the context information is important for online adaption of target localization. To comprehensively leverage the spatial-temporal structure of historical target exemplars and get benefit from the context information, in this work, we present a novel Graph Convolutional Tracking (GCT) method for high-performance visual tracking. Specifically, the GCT jointly incorporates two types of Graph Convolutional Networks (GCNs) into a siamese framework for target appearance modeling. Here, we adopt a spatial-temporal GCN to model the structured representation of historical target exemplars. Furthermore, a context GCN is designed to utilize the context of the current frame to learn adaptive features for target localization. Extensive results on 4 challenging benchmarks show that our GCT method performs favorably against state-of-the-art trackers while running around 50 frames per second.

Zero-shot learning (ZSL) is one of the most promising problems where substantial progress can potentially be achieved through unsupervised learning, due to distributional differences between supervised and zero-shot classes. For this reason, several works investigate the incorporation of discriminative domain adaptation techniques into ZSL, which, however, lead to modest improvements in ZSL accuracy. In contrast, we propose a generative model that can naturally learn from unsupervised examples, and synthesize training examples for unseen classes purely based on their class embeddings, and therefore, reduce the zero-shot learning problem into a supervised classification task. The proposed approach consists of two important components: (i) a conditional Generative Adversarial Network that learns to produce samples that mimic the characteristics of unsupervised data examples, and (ii) the Gradient Matching (GM) loss that measures the quality of the gradient signal obtained from the synthesized examples. Using our GM loss formulation, we enforce the generator to produce examples from which accurate classifiers can be trained. Experimental results on several ZSL benchmark datasets show that our approach leads to significant improvements over the state of the art in generalized zero-shot classification.

Unsupervised domain mapping aims to learn a function GXY to translate domain X to Y in the absence of paired examples. Finding the optimal GXY without paired data is an ill-posed problem, so appropriate constraints are required to obtain reasonable solutions. While some prominent constraints such as cycle consistency and distance preservation successfully constrain the solution space, they overlook the special properties of images that simple geometric transformations do not change the image's semantic structure. Based on this special property, we develop a geometry-consistent generative adversarial network (Gc-GAN), which enables one-sided unsupervised domain mapping. GcGAN takes the original image and its counterpart image transformed by a predefined geometric transformation as inputs and generates two images in the new domain coupled with the corresponding geometry-consistency constraint. The geometry-consistency constraint reduces the space of possible solutions while keep the correct solutions in the search space. Quantitative and qualitative comparisons with the baseline (GAN alone) and the state-of-the-art methods including CycleGAN [66] and DistanceGAN [5] demonstrate the effectiveness of our method.

Surface-based geodesic topology provides strong cues for object semantic analysis and geometric modeling. However, such connectivity information is lost in point clouds. Thus we introduce GeoNet, the first deep learning architecture trained to model the intrinsic structure of surfaces represented as point clouds. To demonstrate the applicability of learned geodesic-aware representations, we propose fusion schemes which use GeoNet in conjunction with other baseline or backbone networks, such as PU-Net and PointNet++, for down-stream point cloud analysis. Our method improves the state-of-the-art on multiple representative tasks that can benefit from understandings of the underlying surface topology, including point upsampling, normal estimation, mesh reconstruction and non-rigid shape classification.

Given an initial recognition model already trained on a set of base classes, the goal of this work is to develop a meta-model for few-shot learning. The meta-model, given as input some novel classes with few training examples per class, must properly adapt the existing recognition model into a new model that can correctly classify in a unified way both the novel and the base classes. To accomplish this goal it must learn to output the appropriate classification weight vectors for those two types of classes. To build our meta-model we make use of two main innovations: we propose the use of a Denoising Autoencoder network (DAE) that (during training) takes as input a set of classification weights corrupted with Gaussian noise and learns to reconstruct the target-discriminative classification weights. In this case, the injected noise on the classification weights serves the role of regularizing the weight generating meta-model. Furthermore, in order to capture the co-dependencies between different classes in a given task instance of our meta-model, we propose to implement the DAE model as a Graph Neural Network (GNN). In order to verify the efficacy of our approach, we extensively evaluate it on ImageNet based few-shot benchmarks and we report state-of-the-art results.

Temporally localizing actions in a video is a fundamental challenge in video understanding. Most existing approaches have often drawn inspiration from image object detection and extended the advances, e.g., SSD and Faster R-CNN, to produce temporal locations of an action in a 1D sequence. Nevertheless, the results can suffer from robustness problem due to the design of predetermined temporal scales, which overlooks the temporal structure of an action and limits the utility on detecting actions with complex variations. In this paper, we propose to address the problem by introducing Gaussian kernels to dynamically optimize temporal scale of each action proposal. Specifically, we present Gaussian Temporal Awareness Networks (GTAN) --- a new architecture that novelly integrates the exploitation of temporal structure into an one-stage action localization framework. Technically, GTAN models the temporal structure through learning a set of Gaussian kernels, each for a cell in the feature maps. Each Gaussian kernel corresponds to a particular interval of an action proposal and a mixture of Gaussian kernels could further characterize action proposals with various length. Moreover, the values in each Gaussian curve reflect the contextual contributions to the localization of an action proposal. Extensive experiments are conducted on both THUMOS14 and ActivityNet v1.3 datasets, and superior results are reported when comparing to state-of-the-art approaches. More remarkably, GTAN achieves 1.9% and 1.1% improvements in mAP on testing set of the two datasets.

Gait, the walking pattern of individuals, is one of the most important biometrics modalities. Most of the existing gait recognition methods take silhouettes or articulated body models as the gait features. These methods suffer from degraded recognition performance when handling confounding variables, such as clothing, carrying and view angle. To remedy this issue, we propose a novel AutoEncoder framework to explicitly disentangle pose and appearance features from RGB imagery and the LSTM-based integration of pose features over time produces the gait feature. In addition, we collect a Frontal-View Gait (FVG) dataset to focus on gait recognition from frontal-view walking, which is a challenging problem since it contains minimal gait cues compared to other views. FVG also includes other important variations,e.g., walking speed, carrying, and clothing. With extensive experiments on CASIA-B, USF and FVG datasets, our method demonstrates superior performance to the-state-of-the-arts quantitatively, the ability of feature disentanglement qualitatively, and promising computational efficiency.

We introduce GQA, a new dataset for real-world visual reasoning and compositional question answering, seeking to address key shortcomings of previous VQA datasets. We have developed a strong and robust question engine that leverages Visual Genome scene graph structures to create 22M diverse reasoning questions, which all come with functional programs that represent their semantics. We use the programs to gain tight control over the answer distribution and present a new tunable smoothing technique to mitigate question biases. Accompanying the dataset is a suite of new metrics that evaluate essential qualities such as consistency, grounding and plausibility. A careful analysis is performed for baselines as well as state-of-the-art models, providing fine-grained results for different question types and topologies. Whereas a blind LSTM obtains a mere 42.1%, and strong VQA models achieve 54.1%, human performance tops at 89.3%, offering ample opportunity for new research to explore. We hope GQA will provide an enabling resource for the next generation of models with enhanced robustness, improved consistency, and deeper semantic understanding of vision and language.

We introduce GFrames, a novel local reference frame (LRF) construction for 3D meshes and point clouds. GFrames are based on the computation of the intrinsic gradient of a scalar field defined on top of the input shape. The resulting tangent vector field defines a repeatable tangent direction of the local frame at each point; importantly, it directly inherits the properties and invariance classes of the underlying scalar function, making it remarkably robust under strong sampling artifacts, vertex noise, as well as non-rigid deformations. Existing local descriptors can directly benefit from our repeatable frames, as we showcase in a selection of 3D vision and shape analysis applications where we demonstrate state-of-the-art performance in a variety of challenging settings.

In the stereo matching task, matching cost aggregation is crucial in both traditional methods and deep neural network models in order to accurately estimate disparities. We propose two novel neural net layers, aimed at capturing local and the whole-image cost dependencies respectively. The first is a semi-global aggregation layer which is a differentiable approximation of the semi-global matching, the second is the local guided aggregation layer which follows a traditional cost filtering strategy to refine thin structures. These two layers can be used to replace the widely used 3D convolutional layer which is computationally costly and memory-consuming as it has cubic computational/memory complexity. In the experiments, we show that nets with a two-layer guided aggregation block easily outperform the state-of-the-art GC-Net which has nineteen 3D convolutional layers. We also train a deep guided aggregation network (GA-Net) which gets better accuracies than state-of-the-art methods on both Scene Flow dataset and KITTI benchmarks.

Visual understanding goes well beyond object recognition. With one glance at an image, we can effortlessly imagine the world beyond the pixels: for instance, we can infer people's actions, goals, and mental states. While this task is easy for humans, it is tremendously difficult for today's vision systems, requiring higher-order cognition and commonsense reasoning about the world. We formalize this task as Visual Commonsense Reasoning. Given a challenging question about an image, a machine must answer correctly and then provide a rationale justifying its answer. Next, we introduce a new dataset, VCR, consisting of 290k multiple choice QA problems derived from 110k movie scenes. The key recipe for generating non-trivial and high-quality problems at scale is Adversarial Matching, a new approach to transform rich annotations into multiple choice questions with minimal bias. Experimental results show that while humans find VCR easy (over 90% accuracy), state-of-the-art vision models struggle ( 45%). To move towards cognition-level understanding, we present a new reasoning engine, Recognition to Cognition Networks (R2C), that models the necessary layered inferences for grounding, contextualization, and reasoning. R2C helps narrow the gap between humans and machines ( 65%); still, the challenge is far from solved, and we provide analysis that suggests avenues for future work.

We propose FineGAN, a novel unsupervised GAN framework, which disentangles the background, object shape, and object appearance to hierarchically generate images of fine-grained object categories. To disentangle the factors without supervision, our key idea is to use information theory to associate each factor to a latent code, and to condition the relationships between the codes in a specific way to induce the desired hierarchy. Through extensive experiments, we show that FineGAN achieves the desired disentanglement to generate realistic and diverse images belonging to fine-grained classes of birds, dogs, and cars. Using FineGAN's automatically learned features, we also cluster real images as a first attempt at solving the novel problem of unsupervised fine-grained object category discovery. Our code/models/demo can be found at https://github.com/kkanshul/finegan

Few-shot learning is an important area of research. Conceptually, humans are readily able to understand new concepts given just a few examples, while in more pragmatic terms, limited-example training situations are common practice. Recent effective approaches to few-shot learning employ a metric-learning framework to learn a feature similarity comparison between a query (test) example, and the few support (training) examples. However, these approaches treat each support class independently from one another, never looking at the entire task as a whole. Because of this, they are constrained to use a single set of features for all possible test-time tasks, which hinders the ability to distinguish the most relevant dimensions for the task at hand. In this work, we introduce a Category Traversal Module that can be inserted as a plug-and-play module into most metric-learning based few-shot learners. This component traverses across the entire support set at once, identifying task-relevant features based on both intra-class commonality and inter-class uniqueness in the feature space. Incorporating our module improves performance considerably (5%-10% relative) over baseline systems on both miniImageNet and tieredImageNet benchmarks, with overall performance competitive with the most recent state-of-the-art systems.

Probabilistic point-set registration methods have been gaining more attention for their robustness to noise, outliers and occlusions. However, these methods tend to be much slower than the popular iterative closest point (ICP) algorithms, which severely limits their usability. In this paper, we contribute a novel probabilistic registration method that achieves state-of-the-art robustness as well as substantially faster computational performance than modern ICP implementations. This is achieved using a rigorous yet computationally-efficient probabilistic formulation. Point-set registration is cast as a maximum likelihood estimation and solved using the EM algorithm. We show that with a simple augmentation, the E step can be formulated as a filtering problem, allowing us to leverage advances in efficient Gaussian filtering methods. We also propose a customized permutohedral filter to improve its performance while retaining sufficient accuracy for our task. Additionally, we present a simple and efficient twist parameterization that generalizes our method to the registration of articulated and deformable objects. For articulated objects, the complexity of our method is almost independent of the Degrees Of Freedom (DOFs), which makes it highly efficient even for high DOF systems. The results demonstrate the proposed method consistently outperforms many competitive baselines on a variety of registration tasks.

Previous works utilized "smaller-norm-less-important" criterion to prune filters with smaller norm values in a convolutional neural network. In this paper, we analyze this norm-based criterion and point out that its effectiveness depends on two requirements that are not always met: (1) the norm deviation of the filters should be large; (2) the minimum norm of the filters should be small. To solve this problem, we propose a novel filter pruning method, namely Filter Pruning via Geometric Median (FPGM), to compress the model regardless of those two requirements. Unlike previous methods, FPGM compresses CNN models by pruning filters with redundancy, rather than those with"relatively less" importance. When applied to two image classification benchmarks, our method validates its usefulness and strengths. Notably, on CIFAR-10, FPGM reduces more than 52% FLOPs on ResNet-110 with even 2.69% relative accuracy improvement. Moreover, on ILSVRC-2012, FPGM reduces more than 42% FLOPs on ResNet-101 without top-5 accuracy drop, which has advanced the state-of-the-art. Code is publicly available on GitHub: https://github.com/he-y/filter-pruning-geometric-median

Traditional recognition methods typically require large, artificially-balanced training classes, while few-shot learning methods are tested on artificially small ones. In contrast to both extremes, real world recognition problems exhibit heavy-tailed class distributions, with cluttered scenes and a mix of coarse and fine-grained class distinctions. We show that prior methods designed for few-shot learning do not work out of the box in these challenging conditions, based on a new "meta-iNat" benchmark. We introduce three parameter-free improvements: (a) better training procedures based on adapting cross-validation to meta-learning, (b) novel architectures that localize objects using limited bounding box annotations before classification, and (c) simple parameter-free expansions of the feature space based on bilinear pooling. Together, these improvements double the accuracy of state-of-the-art models on meta-iNat while generalizing to prior benchmarks, complex neural architectures, and settings with substantial domain shift.

Image captioning is an ambiguous problem, with many suitable captions for an image. To address ambiguity, beam search is the de facto method for sampling multiple captions. However, beam search is computationally expensive and known to produce generic captions. To address this concern, some variational auto-encoder (VAE) and generative adversarial net (GAN) based methods have been proposed. Though diverse, GAN and VAE are less accurate. In this paper, we first predict a meaningful summary of the image, then generate the caption based on that summary. We use part-of-speech as summaries, since our summary should drive caption generation. We achieve the trifecta: (1) High accuracy for the diverse captions as evaluated by standard captioning metrics and user studies; (2) Faster computation of diverse captions compared to beam search and diverse beam search; and (3) High diversity as evaluated by counting novel sentences, distinct n-grams and mutual overlap (i.e., mBleu-4) scores.