CVPR 2019 Papers

We present Scan2CAD, a novel data-driven method that learns to align clean 3D CAD models from a shape database to the noisy and incomplete geometry of a commodity RGB-D scan. For a 3D reconstruction of an indoor scene, our method takes as input a set of CAD models, and predicts a 9DoF pose that aligns each model to the underlying scan geometry. To tackle this problem, we create a new scan-to-CAD alignment dataset based on 1506 ScanNet scans with 97607 annotated keypoint pairs between 14225 CAD models from ShapeNet and their counterpart objects in the scans. Our method selects a set of representative keypoints in a 3D scan for which we find correspondences to the CAD geometry. To this end, we design a novel 3D CNN architecture that learns a joint embedding between real and synthetic objects, and from this predicts a correspondence heatmap. Based on these correspondence heatmaps, we formulate a variational energy minimization that aligns a given set of CAD models to the reconstruction. We evaluate our approach on our newly introduced Scan2CAD benchmark where we outperform both handcrafted feature descriptor as well as state-of-the-art CNN based methods by 21.39%.

Efficient and reliable methods for training of object detectors are in higher demand than ever, and more and more data relevant to the field is becoming available. However, large datasets like Open Images Dataset v4 (OID) are sparsely annotated, and some measure must be taken in order to ensure the training of a reliable detector. In order to take the incompleteness of these datasets into account, one possibility is to use pretrained models to detect the presence of the unverified objects. However, the performance of such a strategy depends largely on the power of the pretrained model. In this study, we propose part-aware sampling, a method that uses human intuition for the hierarchical relation between objects. In terse terms, our method works by making assumptions like "a bounding box for a car should contain a bounding box for a tire". We demonstrate the power of our method on OID and compare the performance against a method based on a pretrained model. Our method also won the first and second place on the public and private test sets of the Google AI Open Images Competition 2018.

In this paper, we propose Spatio-TEmporal Progressive (STEP) action detector--a progressive learning framework for spatio-temporal action detection in videos. Starting from a handful of coarse-scale proposal cuboids, our approach progressively refines the proposals towards actions over a few steps. In this way, high-quality proposals (i.e., adhere to action movements) can be gradually obtained at later steps by leveraging the regression outputs from previous steps. At each step, we adaptively extend the proposals in time to incorporate more related temporal context. Compared to the prior work that performs action detection in one run, our progressive learning framework is able to naturally handle the spatial displacement within action tubes and therefore provides a more effective way for spatio-temporal modeling. We extensively evaluate our approach on UCF101 and AVA, and demonstrate superior detection results. Remarkably, we achieve mAP of 75.0% and 18.6% on the two datasets with 3 progressive steps and using respectively only 11 and 34 initial proposals.

Despite the fact that Second Order Similarity (SOS) has been used with significant success in tasks such as graph matching and clustering, it has not been exploited for learning local descriptors. In this work, we explore the potential of \sos in the field of descriptor learning by building upon the intuition that a positive pair of matching points should exhibit similar distances with respect to other points in the embedding space. Thus, we propose a novel regularization term, named Second Order Similarity Regularization (SOSR), that follows this principle. By incorporating SOSR into training, our learned descriptor achieves state-of-the-art performance on several challenging benchmarks containing distinct tasks ranging from local patch retrieval to structure from motion. Furthermore, by designing a von Mises-Fischer distribution based evaluation method, we link the utilization of the descriptor space to the matching performance, thus demonstrating the effectiveness of our proposed SOSR. Extensive experimental results, empirical evidence, and in-depth analysis are provided, indicating that SOSR can significantly boost the matching performance of the learned descriptor.

We address the problem of semantic correspondence, that is, establishing a dense flow field between images depicting different instances of the same object or scene category. We propose to use images annotated with binary foreground masks and subjected to synthetic geometric deformations to train a convolutional neural network (CNN) for this task. Using these masks as part of the supervisory signal offers a good compromise between semantic flow methods, where the amount of training data is limited by the cost of manually selecting point correspondences, and semantic alignment ones, where the regression of a single global geometric transformation between images may be sensitive to image-specific details such as background clutter. We propose a new CNN architecture, dubbed SFNet, which implements this idea. It leverages a new and differentiable version of the argmax function for end-to-end training, with a loss that combines mask and flow consistency with smoothness terms. Experimental results demonstrate the effectiveness of our approach, which significantly outperforms the state of the art on standard benchmarks.

This paper presents a novel randomized algorithm for robust point cloud registration without correspondences. Most existing registration approaches require a set of putative correspondences obtained by extracting invariant descriptors. However, such descriptors could become unreliable in noisy and contaminated settings. In these settings, methods that directly handle input point sets are preferable. Without correspondences, however, conventional randomized techniques require a very large number of samples in order to reach satisfactory solutions. In this paper, we propose a novel approach to address this problem. In particular, our work enables the use of randomized methods for point cloud registration without the need of putative correspondences. By considering point cloud alignment as a special instance of graph matching and employing an efficient semi-definite relaxation, we propose a novel sampling mechanism, in which the size of the sampled subsets can be larger-than-minimal. Our tight relaxation scheme enables fast rejection of the outliers in the sampled sets, resulting in high quality hypotheses. We conduct extensive experiments to demonstrate that our approach outperforms other state-of-the-art methods. Importantly, our proposed method serves as a generic framework which can be extended to problems with known correspondences.

Dense pixel matching is important for many computer vision tasks such as disparity and flow estimation. We present a robust, unified descriptor network that considers a large context region with high spatial variance. Our network has a very large receptive field and avoids striding layers to maintain spatial resolution. These properties are achieved by creating a novel neural network layer that consists of multiple, parallel, stacked dilated convolutions (SDC). Several of these layers are combined to form our SDC descriptor network. In our experiments, we show that our SDC features outperform state-of-the-art feature descriptors in terms of accuracy and robustness. In addition, we demonstrate the superior performance of SDC in state-of-the-art stereo matching, optical flow and scene flow algorithms on several famous public benchmarks.

Most of the existing subspace clustering (SC) frameworks assume that the noise contaminating the data is generated by an independent and identically distributed (i.i.d.) source, where the Gaussianity is often imposed. Though these assumptions greatly simplify the underlying problems, they do not hold in many real-world applications. For instance, in face clustering, the noise is usually caused by random occlusions, local variations and unconstrained illuminations, which is essentially structural and hence satisfies neither the i.i.d. property nor the Gaussianity. In this work, we propose an independent and piecewise identically distributed (i.p.i.d.) noise model, where the i.i.d. property only holds locally. We demonstrate that the i.p.i.d. model better characterizes the noise encountered in practical scenarios, and accommodates the traditional i.i.d. model as a special case. Assisted by this generalized noise model, we design an information theoretic learning (ITL) framework for robust SC through a novel minimum weighted error entropy (MWEE) criterion. Extensive experimental results show that our proposed SC scheme significantly outperforms the state-of-the-art competing algorithms.

Detection, segmentation and classification of nuclei are fundamental analysis operations in digital pathology. Existing state-of-the-art approaches demand extensive amount of supervised training data from pathologists and may still perform poorly in images from unseen tissue types. We propose an unsupervised approach for histopathology image segmentation that synthesizes heterogeneous sets of training image patches, of every tissue type. Although our synthetic patches are not always of high quality, we harness the motley crew of generated samples through a generally applicable importance sampling method. This proposed approach, for the first time, re-weighs the training loss over synthetic data so that the ideal (unbiased) generalization loss over the true data distribution is minimized. This enables us to use a random polygon generator to synthesize approximate cellular structures (i.e., nuclear masks) for which no real examples are given in many tissue types, and hence, GAN-based methods are not suited. In addition, we propose a hybrid synthesis pipeline that utilizes textures in real histopathology patches and GAN models, to tackle heterogeneity in tissue textures. Compared with existing state-of-the-art supervised models, our approach generalizes significantly better on cancer types without training data. Even in cancer types with training data, our approach achieves the same performance without supervision cost. We release code and segmentation results on over 5000 Whole Slide Images (WSI) in The Cancer Genome Atlas (TCGA) repository, a dataset that would be orders of magnitude larger than what is available today.

Many 3D vision systems localize cameras within a scene using 3D point clouds. Such point clouds are often obtained using structure from motion (SfM), after which the images are discarded to preserve privacy. In this paper, we show, for the first time, that such point clouds retain enough information to reveal scene appearance and compromise privacy. We present a privacy attack that reconstructs color images of the scene from the point cloud. Our method is based on a cascaded U-Net that takes as input, a 2D multichannel image of the points rendered from a specific viewpoint containing point depth and optionally color and SIFT descriptors and outputs a color image of the scene from that viewpoint. Unlike previous feature inversion methods, we deal with highly sparse and irregular 2D point distributions and inputs where many point attributes are missing, namely keypoint orientation and scale, the descriptor image source and the 3D point visibility. We evaluate our attack algorithm on public datasets and analyze the significance of the point cloud attributes. Finally, we show that novel views can also be generated thereby enabling compelling virtual tours of the underlying scene.

Light field cameras are considered to have many potential applications since angular and spatial information is captured simultaneously. However, the limited spatial resolution has brought lots of difficulties in developing related applications and becomes the main bottleneck of light field cameras. In this paper, a learning-based method using residual convolutional networks is proposed to reconstruct light fields with higher spatial resolution. The view images in one light field are first grouped into different image stacks with consistent sub-pixel offsets and fed into different network branches to implicitly learn inherent corresponding relations. The residual information in different spatial directions is then calculated from each branch and further integrated to supplement high-frequency details for the view image. Finally, a flexible solution is proposed to super-resolve entire light field images with various angular resolutions. Experimental results on synthetic and real-world datasets demonstrate that the proposed method outperforms other state-of-the-art methods by a large margin in both visual and numerical evaluations. Furthermore, the proposed method shows good performances in preserving the inherent epipolar property in light field images.

This paper focuses on multi-person action forecasting in videos. More precisely, given a history of H previous frames, the goal is to detect actors and to predict their future actions for the next T frames. Our approach jointly models temporal and spatial interactions among different actors by constructing a recurrent graph, using actor proposals obtained with Faster R-CNN as nodes. Our method learns to select a subset of discriminative relations without requiring explicit supervision, thus enabling us to tackle challenging visual data. We refer to our model as Discriminative Relational Recurrent Network (DRRN). Evaluation of action prediction on AVA demonstrates the effectiveness of our proposed method compared to simpler baselines. Furthermore, we significantly improve performance on the task of early action classification on J-HMDB, from the previous SOTA of 48% to 60%.

Point cloud analysis is very challenging, as the shape implied in irregular points is difficult to capture. In this paper, we propose RS-CNN, namely, Relation-Shape Convolutional Neural Network, which extends regular grid CNN to irregular configuration for point cloud analysis. The key to RS-CNN is learning from relation, i.e., the geometric topology constraint among points. Specifically, the convolutional weight for local point set is forced to learn a high-level relation expression from predefined geometric priors, between a sampled point from this point set and the others. In this way, an inductive local representation with explicit reasoning about the spatial layout of points can be obtained, which leads to much shape awareness and robustness. With this convolution as a basic operator, RS-CNN, a hierarchical architecture can be developed to achieve contextual shape-aware learning for point cloud analysis. Extensive experiments on challenging benchmarks across three tasks verify RS-CNN achieves the state of the arts.

Vision-language navigation (VLN) is the task of navigating an embodied agent to carry out natural language instructions inside real 3D environments. In this paper, we study how to address three critical challenges for this task: the cross-modal grounding, the ill-posed feedback, and the generalization problems. First, we propose a novel Reinforced Cross-Modal Matching (RCM) approach that enforces cross-modal grounding both locally and globally via reinforcement learning (RL). Particularly, a matching critic is used to provide an intrinsic reward to encourage global matching between instructions and trajectories, and a reasoning navigator is employed to perform cross-modal grounding in the local visual scene. Evaluation on a VLN benchmark dataset shows that our RCM model significantly outperforms previous methods by 10% on SPL and achieves the new state-of-the-art performance. To improve the generalizability of the learned policy, we further introduce a Self-Supervised Imitation Learning (SIL) method to explore unseen environments by imitating its own past, good decisions. We demonstrate that SIL can approximate a better and more efficient policy, which tremendously minimizes the success rate performance gap between seen and unseen environments (from 30.7% to 11.7%).

Visual dialog is a challenging vision-language task, which requires the agent to answer multi-round questions about an image. It typically needs to address two major problems: (1) How to answer visually-grounded questions, which is the core challenge in visual question answering (VQA); (2) How to infer the co-reference between questions and the dialog history. An example of visual co-reference is: pronouns (e.g., "they") in the question (e.g., "Are they on or off?") are linked with nouns (e.g., "lamps") appearing in the dialog history (e.g., "How many lamps are there?") and the object grounded in the image. In this work, to resolve the visual co-reference for visual dialog, we propose a novel attention mechanism called Recursive Visual Attention (RvA). Specifically, our dialog agent browses the dialog history until the agent has sufficient confidence in the visual co-reference resolution, and refines the visual attention recursively. The quantitative and qualitative experimental results on the large-scale VisDial v0.9 and v1.0 datasets demonstrate that the proposed RvA not only outperforms the state-of-the-art methods, but also achieves reasonable recursion and interpretable attention maps without additional annotations. The code is available at https://github.com/yuleiniu/rva.

In this paper, we address the large-scale object detection problem with thousands of categories, which poses severe challenges due to long-tail data distributions, heavy occlusions, and class ambiguities. However, the dominant object detection paradigm is limited by treating each object region separately without considering crucial semantic dependencies among objects. In this work, we introduce a novel Reasoning-RCNN to endow any detection networks the capability of adaptive global reasoning over all object regions by exploiting diverse human commonsense knowledge. Instead of only propagating the visual features on the image directly, we evolve the high-level semantic representations of all categories globally to avoid distracted or poor visual features in the image. Specifically, built on feature representations of basic detection network, the proposed network first generates a global semantic pool by collecting the weights of previous classification layer for each category, and then adaptively enhances each object features via attending different semantic contexts in the global semantic pool. Rather than propagating information from all semantic information that may be noisy, our adaptive global reasoning automatically discovers most relative categories for feature evolving. Our Reasoning-RCNN is light-weight and flexible enough to enhance any detection backbone networks, and extensible for integrating any knowledge resources. Solid experiments on object detection benchmarks show the superiority of our Reasoning-RCNN, e.g. achieving around 16% improvement on VisualGenome, 37% on ADE in terms of mAP and 15% improvement on COCO.

We propose a novel model to address the task of Visual Dialog which exhibits complex dialog structures. To obtain a reasonable answer based on the current question and the dialog history, the underlying semantic dependencies between dialog entities are essential. In this paper, we explicitly formalize this task as inference in a graphical model with partially observed nodes and unknown graph structures (relations in dialog). The given dialog entities are viewed as the observed nodes. The answer to a given question is represented by a node with missing value. We first introduce an Expectation Maximization algorithm to infer both the underlying dialog structures and the missing node values (desired answers). Based on this, we proceed to propose a differentiable graph neural network (GNN) solution that approximates this process. Experiment results on the VisDial and VisDial-Q datasets show that our model outperforms comparative methods. It is also observed that our method can infer the underlying dialog structure for better dialog reasoning.

Deep convolutional neural networks trained end-to-end are the state-of-the-art methods to regress dense disparity maps from stereo pairs. These models, however, suffer from a notable decrease in accuracy when exposed to scenarios significantly different from the training set (e.g., real vs synthetic images, etc.). We argue that it is extremely unlikely to gather enough samples to achieve effective training/tuning in any target domain, thus making this setup impractical for many applications. Instead, we propose to perform unsupervised and continuous online adaptation of a deep stereo network, which allows for preserving its accuracy in any environment. However, this strategy is extremely computationally demanding and thus prevents real-time inference. We address this issue introducing a new lightweight, yet effective, deep stereo architecture, Modularly ADaptive Network(MADNet), and developing a Modular ADaptation (MAD) algorithm, which independently trains sub-portions of the network. By deploying MADNet together with MAD we introduce the first real-time self-adaptive deep stereo system enabling competitive performance on heterogeneous datasets. Our code is publicly available at https://github.com/CVLAB-Unibo/Real-time-self-adaptive-deep-stereo.

A well-trained Convolutional Neural Network can easily be pruned without significant loss of performance. This is because of unnecessary overlap in the features captured by the network's filters. Innovations in network architecture such as skip/dense connections and inception units have mitigated this problem to some extent, but these improvements come with increased computation and memory requirements at run-time. We attempt to address this problem from another angle - not by changing the network structure but by altering the training method. We show that by temporarily pruning and then restoring a subset of the model's filters, and repeating this process cyclically, overlap in the learned features is reduced, producing improved generalization. We show that the existing model-pruning criteria are not optimal for selecting filters to prune in this context, and introduce inter-filter orthogonality as the ranking criteria to determine under-expressive filters. Our method is applicable both to vanilla convolutional networks and more complex modern architectures, and improves the performance across a variety of tasks, especially when applied to smaller networks.