Data augmentation has been widely used to improve generalizability of machine learning models. However, comparatively little work studies data augmentation for graphs. This is largely due to the complex, non-Euclidean structure of graphs, which limits possible manipulation operations. Augmentation operations commonly used in vision and language have no analogs for graphs. Our work studies graph data augmentation for graph neural networks (GNNs) in the context of improving semi-supervised node-classification. We discuss practical and theoretical motivations, considerations and strategies for graph data augmentation. Our work shows that neural edge predictors can effectively encode class-homophilic structure to promote intra-class edges and demote inter-class edges in given graph structure, and our main contribution introduces the GAug graph data augmentation framework, which leverages these insights to improve performance in GNN-based node classification via edge prediction. Extensive experiments on multiple benchmarks show that augmentation via GAug improves performance across GNN architectures and datasets.

Generative Adversarial Networks (GANs) have achieved huge success in generating high-fidelity images, however, they suffer from low efficiency due to tremendous computational cost and bulky memory usage. Recent efforts on compression GANs show noticeable progress in obtaining smaller generators by sacrificing image quality or involving a time-consuming searching process. In this work, we aim to address these issues by introducing a teacher network that provides a search space in which efficient network architectures can be found, in addition to performing knowledge distillation. First, we revisit the search space of generative models, introducing an inception-based residual block into generators. Second, to achieve target computation cost, we propose a one-step pruning algorithm that searches a student architecture from the teacher model and substantially reduces searching cost. It requires no ℓ1sparsity regularization and its associated hyper-parameters, simplifying the training procedure. Finally, we propose to distill knowledge through maximizing feature similarity between teacher and student via an index named Global Kernel Alignment (GKA). Our compressed networks achieve similar or even better image fidelity (FID, mIoU) than the original models with much-reduced computational cost, e.g., MACs. Code will be released at https://github.com/snap-research/CAT.

Human motion retargeting aims to transfer the motion of one person in a "driving" video or set of images to another person. Existing efforts leverage a long training video from each target person to train a subject-specific motion transfer model. However, the scalability of such methods is limited, as each model can only generate videos for the given target subject, and such training videos are labor-intensive to acquire and process. Few-shot motion transfer techniques, which only require one or a few images from a target, have recently drawn considerable attention. Methods addressing this task generally use either 2D or explicit 3D representations to transfer motion, and in doing so, sacrifice either accurate geometric modeling or the flexibility of an end-to-end learned representation. Inspired by the Transformable Bottleneck Network, which renders novel views and manipulations of rigid objects, we propose an approach based on an implicit volumetric representation of the image content, which can then be spatially manipulated using volumetric flow fields. We address the challenging question of how to aggregate information across different body poses, learning flow fields that allow for combining content from the appropriate regions of input images of highly non-rigid human subjects performing complex motions into a single implicit volumetric representation. This allows us to learn our 3D representation solely from videos of moving people. Armed with both 3D object understanding and end-to-end learned rendering, this categorically novel representation delivers state-of-the-art image generation quality, as shown by our quantitative and qualitative evaluations.

We propose novel motion representations for animating articulated objects consisting of distinct parts. In a completely unsupervised manner, our method identifies object parts, tracks them in a driving video, and infers their motions by considering their principal axes. In contrast to the previous keypoint-based works, our method extracts meaningful and consistent regions, describing locations, shape, and pose. The regions correspond to semantically relevant and distinct object parts, that are more easily detected in frames of the driving video. To force decoupling of foreground from background, we model non-object related global motion with an additional affine transformation. To facilitate animation and prevent the leakage of the shape of the driving object, we disentangle shape and pose of objects in the region space. Our model1can animate a variety of objects, surpassing previous methods by a large margin on existing benchmarks. We present a challenging new benchmark with high-resolution videos and show that the improvement is particularly pronounced when articulated objects are considered, reaching 96.6% user preference vs. the state of the art.

Full-motion cost volumes play a central role in current state-of-the-art optical flow methods. However, constructed using simple feature correlations, they lack the ability to encapsulate prior, or even non-local knowledge. This creates artifacts in poorly constrained ambiguous regions, such as occluded and textureless areas. We propose a separable cost volume module, a drop-in replacement to correlation cost volumes, that uses non-local aggregation layers to exploit global context cues and prior knowledge, in order to disambiguate motions in these regions. Our method leads both the now standard Sintel and KITTI optical flow benchmarks in terms of accuracy, and is also shown to generalize better from synthetic to real data.

We propose a novel approach to performing fine-grained 3D manipulation of image content via a convolutional neural network, which we call the Transformable Bottleneck Network (TBN). It applies given spatial transformations directly to a volumetric bottleneck within our encoder-bottleneck-decoder architecture. Multi-view supervision encourages the network to learn to spatially disentangle the feature space within the bottleneck. The resulting spatial structure can be manipulated with arbitrary spatial transformations. We demonstrate the efficacy of TBNs for novel view synthesis, achieving state-of-the-art results on a challenging benchmark. We demonstrate that the bottlenecks produced by networks trained for this task contain meaningful spatial structure that allows us to intuitively perform a variety of image manipulations in 3D, well beyond the rigid transformations seen during training. These manipulations include non-uniform scaling, non-rigid warping, and combining content from different images. Finally, we extract explicit 3D structure from the bottleneck, performing impressive 3D reconstruction from a single input image.

Despite the popularity of real-time monocular face tracking systems in many successful applications, one overlooked problem with these systems is rigid instability. It occurs when the input facial motion can be explained by either head pose change or facial expression change, creating ambiguities that often lead to jittery and unstable rigid head poses under large expressions. Existing rigid stabilization methods either employ a heavy anatomically-motivated approach that are unsuitable for real-time applications, or utilize heuristic-based rules that can be problematic under certain expressions. We propose the first rigid stabilization method for real-time monocular face tracking using a dynamic rigidity prior learned from realistic datasets. The prior is defined on a region-based face model and provides dynamic region-based adaptivity for rigid pose optimization during real-time performance. We introduce an effective offline training scheme to learn the dynamic rigidity prior by optimizing the convergence of the rigid pose optimization to the ground-truth poses in the training data. Our real-time face tracking system is an optimization framework that alternates between rigid pose optimization and expression optimization. To ensure tracking accuracy, we combine both robust, drift-free facial landmarks and dense optical flow into the optimization objectives. We evaluate our system extensively against state-of-the-art monocular face tracking systems and achieve significant improvement in tracking accuracy on the high-quality face tracking benchmark. Our system can improve facial-performance-based applications such as facial animation retargeting and virtual face makeup with accurate expression and stable pose. We further validate the dynamic rigidity prior by comparing it against other variants on the tracking accuracy.

We present the application of 3D reconstruction technology to the inspection and decommissioning work at the damaged Fukushima Daiichi nuclear power station in Japan. We discuss the challenges of this project, such as the difficult image capture conditions (including under water), required use of limited imaging hardware, and capture by personnel inexperienced in 3D reconstruction. We present an overview of the system developed for this project, a real-time reconstruction pipeline with robust camera pose estimation, low-latency probabilistic dense depth estimation and a novel descriptor for point cloud alignment - the Co-occurrence Histogram of Angle and Distance (CHAD). We discuss the modifications required to standard algorithms in order to perform reliably in such a scenario. As well as quantitative evaluations of these components on existing datasets, we show qualitative 3D reconstruction results of debris from the damaged plant and its spent fuel pool. Such results have enabled planning of the critical process of debris removal, without the harmful requirement of extensive human presence on site.

In applying the Hough transform to the problem of 3D shape recognition and registration, we develop two new and powerful improvements to this popular inference method. The first, intrinsic Hough, solves the problem of exponential memory requirements of the standard Hough transform by exploiting the sparsity of the Hough space. The second, minimum-entropy Hough, explains away incorrect votes, substantially reducing the number of modes in the posterior distribution of class and pose, and improving precision. Our experiments demonstrate that these contributions make the Hough transform not only tractable but also highly accurate for our example application. Both contributions can be applied to other tasks that already use the standard Hough transform.

This paper presents a method for vote-based 3D shape recognition and registration, in particular using mean shift on 3D pose votes in the space of direct similarity transforms for the first time. We introduce a new distance between poses in this space-the SRT distance. It is left-invariant, unlike Euclidean distance, and has a unique, closed-form mean, in contrast to Riemannian distance, so is fast to compute. We demonstrate improved performance over the state of the art in both recognition and registration on a real and challenging dataset, by comparing our distance with others in a mean shift framework, as well as with the commonly used Hough voting approach.

In recent years the Markov Random Field (MRF) has become the de facto probabilistic model for low-level vision applications. However, in a maximum a posteriori (MAP) framework, MRFs inherently encourage delta function marginal statistics. By contrast, many low-level vision problems have heavy tailed marginal statistics, making the MRF model unsuitable. In this paper we introduce a more general Marginal Probability Field (MPF), of which the MRF is a special, linear case, and show that convex energy MPFs can be used to encourage arbitrary marginal statistics. We introduce a flexible, extensible framework for effectively optimizing the resulting NP-hard MAP problem, based around dual-decomposition and a modified min-cost flow algorithm, and which achieves global optimality in some instances. We use a range of applications, including image denoising and texture synthesis, to demonstrate the benefits of this class of MPF over MRFs.

Second-order priors on the smoothness of 3D surfaces are a better model of typical scenes than first-order priors. However, stereo reconstruction using global inference algorithms, such as graph cuts, has not been able to incorporate second-order priors because the triple cliques needed to express them yield intractable (nonsubmodular) optimization problems. This paper shows that inference with triple cliques can be effectively performed. Our optimization strategy is a development of recent extensions to alpha -- expansion, based on the "QPBO" algorithm. The strategy is to repeatedly merge proposal depth maps using a novel extension of QPBO. Proposal depth maps can come from any source, for example, frontoparallel planes as in alpha-expansion, or indeed any existing stereo algorithm, with arbitrary parameter settings.

Second-order priors on the smoothness of 3D surfaces are a better model of typical scenes than first-order priors. However, stereo reconstruction using global inference algorithms, such as graph-cuts, has not been able to incorporate second-order priors because the triple cliques needed to express them yield intractable (non-submodular) optimization problems. This paper shows that inference with triple cliques can be effectively optimized. Our optimization strategy is a development of recent extensions to a-expansion, based on the "QPBO" algorithm [5, 14, 26]. The strategy is to repeatedly merge proposal depth maps using a novel extension of QPBO. Proposal depth maps can come from any source, for example fronto-parallel planes as in a-expansion, or indeed any existing stereo algorithm, with arbitrary parameter settings. Experimental results demonstrate the usefulness of the second-order prior and the efficacy of our optimization framework. An implementation of our stereo framework is available online [34].

We show that application of modern multiview stereo techniques to the newview synthesis (NVS) problem introduces a number of non-trivial complexities. By simultaneously solving for the colour and depth of the new-view pixels we can eliminate the visual artefacts that conventional NVS-via-stereo suffers. The global occlusion reasoning which has led to considerable improvements in recent stereo algorithms can easily be included in the new algorithm, using a recently improved graph-cut-based optimizer for general multi-label conditional random fields (CRFs). However, the CRF priors that are important to success in stereo cannot be easily applied if the reconstruction is to be computed in the reference frame of the novel view. We address this problem by extending recent work on the fast optimization of texture priors in NVS to model the image edge structure, yielding a synthesis of the two approaches which yields good results on difficult image sequences. 1

New-view synthesis (NVS) using texture priors (as opposed to surface-smoothness priors) can yield high quality results, but the standard formulation is in terms of large-clique Markov random fields (MRFs). Only local optimization methods such as iterated conditional modes, which are prone to fall into local minima close to the initial estimate, are practical for solving these problems. In this paper we replace the large-clique energies with pairwise potentials, by restricting the patch dictionary for each clique to image regions suitable for that clique. This enables for the first time the use of a global optimization method, such as tree-reweighted message passing, to solve the NVS problem with image-based priors. We employ a robust, truncated quadratic kernel to reject outliers caused by occlusions, specularities and moving objects, within our global optimization. Because the MRF optimization is thus fast, computing the unary potentials becomes the new performance bottleneck. An additional contribution of this paper is a novel, fast method for enumerating color modes of the per-pixel unary potentials, despite the non-convex nature of our robust kernel. We compare the results of our technique with other rendering methods, and discuss the relative merits and flaws of regularizing color, and of local versus global dictionaries.

Image priors for novel view synthesis have traditionally been non-parametric models based on large libraries of image patch exemplars, producing highquality results but making inference very slow. Recently a parametric framework, called Fields of Experts, has been proposed for image restoration that promises to speed up inference dramatically. In this paper we apply Fields of Experts for the first time to the problem of novel view synthesis, posed as a Markov random field labelling problem with very large cliques. Additionally, we introduce to computer vision for the first time a new optimization algorithm from statistical physics which reaches better minima than the ICM and simulated annealing algorithms to which such large-clique problems have previously been restricted. 1