ICCV 2019 论文列表

Domain shift is unavoidable in real-world applications of object detection. For example, in self-driving cars, the target domain consists of unconstrained road environments which cannot all possibly be observed in training data. Similarly, in surveillance applications sufficiently representative training data may be lacking due to privacy regulations. In this paper, we address the domain adaptation problem from the perspective of robust learning and show that the problem may be formulated as training with noisy labels. We propose a robust object detection framework that is resilient to noise in bounding box class labels, locations and size annotations. To adapt to the domain shift, the model is trained on the target domain using a set of noisy object bounding boxes that are obtained by a detection model trained only in the source domain. We evaluate the accuracy of our approach in various source/target domain pairs and demonstrate that the model significantly improves the state-of-the-art on multiple domain adaptation scenarios on the SIM10K, Cityscapes and KITTI datasets.

Unsupervised cross-domain person re-identification (Re-ID) faces two key issues. One is the data distribution discrepancy between source and target domains, and the other is the lack of discriminative information in target domain. From the perspective of representation learning, this paper proposes a novel end-to-end deep domain adaptation framework to address them. For the first issue, we highlight the presence of camera-level sub-domains as a unique characteristic in person Re-ID, and develop a "camera-aware" domain adaptation method via adversarial learning. With this method, the learned representation reduces distribution discrepancy not only between source and target domains but also across all cameras. For the second issue, we exploit the temporal continuity in each camera of target domain to create discriminative information. This is implemented by dynamically generating online triplets within each batch, in order to maximally take advantage of the steadily improved representation in training process. Together, the above two methods give rise to a new unsupervised domain adaptation framework for person Re-ID. Extensive experiments and ablation studies conducted on benchmark datasets demonstrate its superiority and interesting properties.

Dramatic advances in generative models have resulted in near photographic quality for artificially rendered faces, animals and other objects in the natural world. In spite of such advances, a higher level understanding of vision and imagery does not arise from exhaustively modeling an object, but instead identifying higher-level attributes that best summarize the aspects of an object. In this work we attempt to model the drawing process of fonts by building sequential generative models of vector graphics. This model has the benefit of providing a scale-invariant representation for imagery whose latent representation may be systematically manipulated and exploited to perform style propagation. We demonstrate these results on a large dataset of fonts crawled from the web and highlight how such a model captures the statistical dependencies and richness of this dataset. We envision that our model can find use as a tool for graphic designers to facilitate font design.

Deep neural networks have recently achieved tremendous success in image classification. Recent studies have however shown that they are easily misled into incorrect classification decisions by adversarial examples. Adversaries can even craft attacks by querying the model in black-box settings, where no information about the model is released except its final decision. Such decision-based attacks usually require lots of queries, while real-world image recognition systems might actually restrict the number of queries. In this paper, we propose qFool, a novel decision-based attack algorithm that can generate adversarial examples using a small number of queries. The qFool method can drastically reduce the number of queries compared to previous decision-based attacks while reaching the same quality of adversarial examples. We also enhance our method by constraining adversarial perturbations in low-frequency subspace, which can make qFool even more computationally efficient. Altogether, we manage to fool commercial image recognition systems with a small number of queries, which demonstrates the actual effectiveness of our new algorithm in practice.

Highly accurate 3D volumetric reconstruction is still an open research topic where the main difficulty is usually related to merging some rough estimations with high frequency details. One of the most promising methods is the fusion between multi-view stereo and photometric stereo images. Beside the intrinsic difficulties that multi-view stereo and photometric stereo in order to work reliably, supplementary problems arise when considered together. In this work, we present a volumetric approach to the multi-view photometric stereo problem. The key point of our method is the signed distance field parameterisation and its relation to the surface normal. This is exploited in order to obtain a linear partial differential equation which is solved in a variational framework, that combines multiple images from multiple points of view in a single system. In addition, the volumetric approach is naturally implemented on an octree, which allows for fast ray-tracing that reliably alleviates occlusions and cast shadows. Our approach is evaluated on synthetic and real data-sets and achieves state-of-the-art results.

Average precision (AP) is a widely used metric to evaluate detection accuracy of image and video object detectors. In this paper, we analyze the object detection from video and point out that mAP alone is not sufficient to capture the temporal nature of video object detection. To tackle this problem, we propose a comprehensive metric, Average Delay (AD), to measure and compare detection delay. To facilitate delay evaluation, we carefully select a subset of ImageNet VID, which we name as ImageNet VIDT with an emphasis on complex trajectories. By extensively evaluating a wide range of detectors on VIDT, we show that most methods drastically increase the detection delay but still preserve mAP well. In other words, mAP is not sensitive enough to reflect the temporal characteristics of a video object detector. Our results suggest that video object detection methods should be evaluated with a delay metric, particularly for latency-critical applications such as autonomous vehicle perception.

This paper presents a novel feature-based method that is built upon a convolutional neural network (CNN) to learn the deep representation for multimodal retinal image registration. We coined the algorithm deep step patterns, in short DeepSPa. Most existing deep learning based methods require a set of manually labeled training data with known corresponding spatial transformations, which limits the size of training datasets. By contrast, our method is fully automatic and scale well to different image modalities with no human intervention. We generate feature classes from simple step patterns within patches of connecting edges formed by vascular junctions in multiple retinal imaging modalities. We leverage CNN to learn and optimize the input patches to be used for image registration. Spatial transformations are estimated based on the output possibility of the fully connected layer of CNN for a pair of images. One of the key advantages of the proposed algorithm is its robustness to non-linear intensity changes, which widely exist on retinal images due to the difference of acquisition modalities. We validate our algorithm on extensive challenging datasets comprising poor quality multimodal retinal images which are adversely affected by pathologies (diseases), speckle noise and low resolutions. The experimental results demonstrate the robustness and accuracy over state-of-the-art multimodal image registration algorithms.

What if we could interpret the cognitive state of a user while experiencing a virtual reality (VR) and estimate the cognitive state from a visual stimulus? In this paper, we address the above question by developing an electroencephalography (EEG) driven VR cybersickness prediction model. The EEG data has been widely utilized to learn the cognitive representation of brain activity. In the first stage, to fully exploit the advantages of the EEG data, it is transformed into the multi-channel spectrogram which enables to account for the correlation of spectral and temporal coefficient. Then, a convolutional neural network (CNN) is applied to encode the cognitive representation of the EEG spectrogram. In the second stage, we train a cybersickness prediction model on the VR video sequence by designing a Recurrent Neural Network (RNN). Here, the encoded cognitive representation is transferred to the model to train the visual and cognitive features for cybersickness prediction. Through the proposed framework, it is possible to predict the cybersickness level that reflects brain activity automatically. We use 8-channels EEG data to record brain activity while more than 200 subjects experience 44 different VR contents. After rigorous training, we demonstrate that the proposed framework reliably estimates cognitive states without the EEG data. Furthermore, it achieves state-of-the-art performance comparing to existing VR cybersickness prediction models.

Collections of images under a single, uncontrolled illumination have enabled the rapid advancement of core computer vision tasks like classification, detection, and segmentation. But even with modern learning techniques, many inverse problems involving lighting and material understanding remain too severely ill-posed to be solved with single-illumination datasets. The data simply does not contain the necessary supervisory signals. Multi-illumination datasets are notoriously hard to capture, so the data is typically collected at small scale, in controlled environments, either using multiple light sources, or robotic gantries. This leads to image collections that are not representative of the variety and complexity of real world scenes. We introduce a new multi-illumination dataset of more than 1000 real scenes, each captured in high dynamic range and high resolution, under 25 lighting conditions. We demonstrate the richness of this dataset by training state-of-the-art models for three challenging applications: single-image illumination estimation, image relighting, and mixed-illuminant white balance.

We investigate the design aspects of feature distillation methods achieving network compression and propose a novel feature distillation method in which the distillation loss is designed to make a synergy among various aspects: teacher transform, student transform, distillation feature position and distance function. Our proposed distillation loss includes a feature transform with a newly designed margin ReLU, a new distillation feature position, and a partial L2 distance function to skip redundant information giving adverse effects to the compression of student. In ImageNet, our proposed method achieves 21.65% of top-1 error with ResNet50, which outperforms the performance of the teacher network, ResNet152. Our proposed method is evaluated on various tasks such as image classification, object detection and semantic segmentation and achieves a significant performance improvement in all tasks. The code is available at project page.

Universal style transfer tries to explicitly minimize the losses in feature space, thus it does not require training on any pre-defined styles. It usually uses different layers of VGG network as the encoders and trains several decoders to invert the features into images. Therefore, the effect of style transfer is achieved by feature transform. Although plenty of methods have been proposed, a theoretical analysis of feature transform is still missing. In this paper, we first propose a novel interpretation by treating it as the optimal transport problem. Then, we demonstrate the relations of our formulation with former works like Adaptive Instance Normalization (AdaIN) and Whitening and Coloring Transform (WCT). Finally, we derive a closed-form solution named Optimal Style Transfer (OST) under our formulation by additionally considering the content loss of Gatys. Comparatively, our solution can preserve better structure and achieve visually pleasing results. It is simple yet effective and we demonstrate its advantages both quantitatively and qualitatively. Besides, we hope our theoretical analysis can inspire future works in neural style transfer.

Neural network compression is an important step for deploying neural networks where speed is of high importance, or on devices with limited memory. It is necessary to tune compression parameters in order to achieve the desired trade-off between size and performance. This is often done by optimizing the loss on a validation set of data, which should be large enough to approximate the true risk and therefore yield sufficient generalization ability. However, using a full validation set can be computationally expensive. In this work, we develop a general Bayesian optimization framework for optimizing functions that are computed based on U-statistics. We propagate Gaussian uncertainties from the statistics through the Bayesian optimization framework yielding a method that gives a probabilistic approximation certificate of the result. We then apply this to parameter selection in neural network compression. Compression objectives that can be written as U-statistics are typically based on empirical risk and knowledge distillation for deep discriminative models. We demonstrate our method on VGG and ResNet models, and the resulting system can find optimal compression parameters for relatively high-dimensional parametrizations in a matter of minutes on a standard desktop machine, orders of magnitude faster than competing methods.

Generating grasp poses is a crucial component for any robot object manipulation task. In this work, we formulate the problem of grasp generation as sampling a set of grasps using a variational autoencoder and assess and refine the sampled grasps using a grasp evaluator model. Both Grasp Sampler and Grasp Refinement networks take 3D point clouds observed by a depth camera as input. We evaluate our approach in simulation and real-world robot experiments. Our approach achieves 88% success rate on various commonly used objects with diverse appearances, scales, and weights. Our model is trained purely in simulation and works in the real-world without any extra steps.

Recent advances in 3D human shape estimation build upon parametric representations that model very well the shape of the naked body, but are not appropriate to represent the clothing geometry. In this paper, we present an approach to model dressed humans and predict their geometry from single images. We contribute in three fundamental aspects of the problem, namely, a new dataset, a novel shape parameterization algorithm and an end-to-end deep generative network for predicting shape. First, we present 3DPeople, a large-scale synthetic dataset with 2 Million photo-realistic images of 80 subjects performing 70 activities and wearing diverse outfits. Besides providing textured 3D meshes for clothes and body we annotated the dataset with segmentation masks, skeletons, depth, normal maps and optical flow. All this together makes 3DPeople suitable for a plethora of tasks. We then represent the 3D shapes using 2D geometry images. To build these images we propose a novel spherical area-preserving parameterization algorithm based on the optimal mass transportation method. We show this approach to improve existing spherical maps which tend to shrink the elongated parts of the full body models such as the arms and legs, making the geometry images incomplete. Finally, we design a multi-resolution deep generative network that, given an input image of a dressed human, predicts his/her geometry image (and thus the clothed body shape) in an end-to-end manner. We obtain very promising results in jointly capturing body pose and clothing shape, both for synthetic validation and on the wild images.

We propose an approach to predict the 3D shape and pose for the objects present in a scene. Existing learning based methods that pursue this goal make independent predictions per object, and do not leverage the relationships amongst them. We argue that reasoning about these relationships is crucial, and present an approach to incorporate these in a 3D prediction framework. In addition to independent per-object predictions, we predict pairwise relations in the form of relative 3D pose, and demonstrate that these can be easily incorporated to improve object level estimates. We report performance across different datasets (SUNCG, NYUv2), and show that our approach significantly improves over independent prediction approaches while also outperforming alternate implicit reasoning methods.

We introduce a network that directly predicts the 3D layout of lanes in a road scene from a single image. This work marks a first attempt to address this task with on-board sensing without assuming a known constant lane width or relying on pre-mapped environments. Our network architecture, 3D-LaneNet, applies two new concepts: intra-network inverse-perspective mapping (IPM) and anchor-based lane representation. The intra-network IPM projection facilitates a dual-representation information flow in both regular image-view and top-view. An anchor-per-column output representation enables our end-to-end approach which replaces common heuristics such as clustering and outlier rejection, casting lane estimation as an object detection problem. In addition, our approach explicitly handles complex situations such as lane merges and splits. Results are shown on two new 3D lane datasets, a synthetic and a real one. For comparison with existing methods, we test our approach on the image-only tuSimple lane detection benchmark, achieving performance competitive with state-of-the-art.

We tackle the problem of automatically reconstructing a complete 3D model of a scene from a single RGB image. This challenging task requires inferring the shape of both visible and occluded surfaces. Our approach utilizes viewer-centered, multi-layer representation of scene geometry adapted from recent methods for single object shape completion. To improve the accuracy of view-centered representations for complex scenes, we introduce a novel "Epipolar Feature Transformer" that transfers convolutional network features from an input view to other virtual camera viewpoints, and thus better covers the 3D scene geometry. Unlike existing approaches that first detect and localize objects in 3D, and then infer object shape using category-specific models, our approach is fully convolutional, end-to-end differentiable, and avoids the resolution and memory limitations of voxel representations. We demonstrate the advantages of multi-layer depth representations and epipolar feature transformers on the reconstruction of a large database of indoor scenes.

A comprehensive semantic understanding of a scene is important for many applications - but in what space should diverse semantic information (e.g., objects, scene categories, material types, 3D shapes, etc.) be grounded and what should be its structure? Aspiring to have one unified structure that hosts diverse types of semantics, we follow the Scene Graph paradigm in 3D, generating a 3D Scene Graph. Given a 3D mesh and registered panoramic images, we construct a graph that spans the entire building and includes semantics on objects (e.g., class, material, shape and other attributes), rooms (e.g., function, illumination type, etc.) and cameras (e.g., location, etc.), as well as the relationships among these entities. However, this process is prohibitively labor heavy if done manually. To alleviate this we devise a semi-automatic framework that employs existing detection methods and enhances them using two main constraints: I. framing of query images sampled on panoramas to maximize the performance of 2D detectors, and II. multi-view consistency enforcement across 2D detections that originate in different camera locations.

In this paper, we propose a novel generative adversarial network (GAN) for 3D point clouds generation, which is called tree-GAN. To achieve state-of-the-art performance for multi-class 3D point cloud generation, a tree-structured graph convolution network (TreeGCN) is introduced as a generator for tree-GAN. Because TreeGCN performs graph convolutions within a tree, it can use ancestor information to boost the representation power for features. To evaluate GANs for 3D point clouds accurately, we develop a novel evaluation metric called Frechet point cloud distance (FPD). Experimental results demonstrate that the proposed tree-GAN outperforms state-of-the-art GANs in terms of both conventional metrics and FPD, and can generate point clouds for different semantic parts without prior knowledge.