ICCV 2019 Papers

This paper addresses the problem of deformable surface tracking from monocular images. Specifically, we propose a graph-based approach that effectively explores the structure information of the surface to enhance tracking performance. Our approach solves simultaneously for feature correspondence, outlier rejection and shape reconstruction by optimizing a single objective function, which is defined by means of pairwise projection errors between graph structures instead of unary projection errors between matched points. Furthermore, an efficient matching algorithm is developed based on soft matching relaxation. For evaluation, our approach is extensively compared to state-of-the-art algorithms on a standard dataset of occluded surfaces, as well as a newly compiled dataset of different surfaces with rich, weak or repetitive texture. Experimental results reveal that our approach achieves robust tracking results for surfaces with different types of texture, and outperforms other algorithms in both accuracy and efficiency.

Classifiers such as deep neural networks have been shown to be vulnerable against adversarial perturbations on problems with high-dimensional input space. While adversarial training improves the robustness of image classifiers against such adversarial perturbations, it leaves them sensitive to perturbations on a non-negligible fraction of the inputs. In this work, we show that adversarial training is more effective in preventing universal perturbations, where the same perturbation needs to fool a classifier on many inputs. Moreover, we investigate the trade-off between robustness against universal perturbations and performance on unperturbed data and propose an extension of adversarial training that handles this trade-off more gracefully. We present results for image classification and semantic segmentation to showcase that universal perturbations that fool a model hardened with adversarial training become clearly perceptible and show patterns of the target scene.

We present DeepVCP - a novel end-to-end learning-based 3D point cloud registration framework that achieves comparable registration accuracy to prior state-of-the-art geometric methods. Different from other keypoint based methods where a RANSAC procedure is usually needed, we implement the use of various deep neural network structures to establish an end-to-end trainable network. Our keypoint detector is trained through this end-to-end structure and enables the system to avoid the interference of dynamic objects, leverages the help of sufficiently salient features on stationary objects, and as a result, achieves high robustness. Rather than searching the corresponding points among existing points, the key contribution is that we innovatively generate them based on learned matching probabilities among a group of candidates, which can boost the registration accuracy. We comprehensively validate the effectiveness of our approach using both the KITTI dataset and the Apollo-SouthBay dataset. Results demonstrate that our method achieves comparable registration accuracy and runtime efficiency to the state-of-the-art geometry-based methods, but with higher robustness to inaccurate initial poses. Detailed ablation and visualization analysis are included to further illustrate the behavior and insights of our network. The low registration error and high robustness of our method make it attractive to the substantial applications relying on the point cloud registration task.

Our goal is to significantly speed up the runtime of current state-of-the-art stereo algorithms to enable real-time inference. Towards this goal, we developed a differentiable PatchMatch module that allows us to discard most disparities without requiring full cost volume evaluation. We then exploit this representation to learn which range to prune for each pixel. By progressively reducing the search space and effectively propagating such information, we are able to efficiently compute the cost volume for high likelihood hypotheses and achieve savings in both memory and computation.Finally, an image guided refinement module is exploited to further improve the performance. Since all our components are differentiable, the full network can be trained end-to-end. Our experiments show that our method achieves competitive results on KITTI and SceneFlow datasets while running in real-time at 62ms.

Snapshot compressive imaging (SCI) systems have been developed to capture high-dimensional (> 3) signals using low-dimensional off-the-shelf sensors, i.e., mapping multiple video frames into a single measurement frame. One key module of a SCI system is an accurate decoder that recovers the original video frames. However, existing model-based decoding algorithms require exhaustive parameter tuning with prior knowledge and cannot support practical applications due to the extremely long running time. In this paper, we propose a deep tensor ADMM-Net for video SCI systems that provides high-quality decoding in seconds. Firstly, we start with a standard tensor ADMM algorithm, unfold its inference iterations into a layer-wise structure, and design a deep neural network based on tensor operations. Secondly, instead of relying on a pre-specified sparse representation domain, the network learns the domain of low-rank tensor through stochastic gradient descent. It is worth noting that the proposed deep tensor ADMM-Net has potentially mathematical interpretations. On public video data, the simulation results show the proposed method achieves average 0.8 ~ 2.5 dB improvement in PSNR and 0.07 ~ 0.1 in SSIM, and 1500x~ 3600 xspeedups over the state-of-the-art methods. On real data captured by SCI cameras, the experimental results show comparable visual results with the state-of-the-art methods but in much shorter running time.

Recently, a series of deep supervised hashing methods were proposed for binary code learning. However, due to the high computation cost and the limited hardware's memory, these methods will first select a subset from the training set, and then form a mini-batch data to update the network in each iteration. Therefore, the remaining labeled data cannot be fully utilized and the model cannot directly obtain the binary codes of the entire training set for retrieval. To address these problems, this paper proposes an interesting regularized deep model to seamlessly integrate the advantages of deep hashing and efficient binary code learning by using the anchor graph. As such, the deep features and label matrix can be jointly used to optimize the binary codes, and the network can obtain more discriminative feedback from the linear combinations of the learned bits. Moreover, we also reveal the algorithm mechanism and its computation essence. Experiments on three large-scale datasets indicate that the proposed method achieves better retrieval performance with less training time compared to previous deep hashing methods.

Conventional physically-based methods for relighting portrait images need to solve an inverse rendering problem, estimating face geometry, reflectance and lighting. However, the inaccurate estimation of face components can cause strong artifacts in relighting, leading to unsatisfactory results. In this work, we apply a physically-based portrait relighting method to generate a large scale, high quality, "in the wild" portrait relighting dataset (DPR). A deep Convolutional Neural Network (CNN) is then trained using this dataset to generate a relit portrait image by using a source image and a target lighting as input. The training procedure regularizes the generated results, removing the artifacts caused by physically-based relighting methods. A GAN loss is further applied to improve the quality of the relit portrait image. Our trained network can relight portrait images with resolutions as high as 1024 x 1024. We evaluate the proposed method on the proposed DPR datset, Flickr portrait dataset and Multi-PIE dataset both qualitatively and quantitatively. Our experiments demonstrate that the proposed method achieves state-of-the-art results. Please refer to https://zhhoper.github.io/dpr.html for dataset and code.

ConvNets achieve good results when training from clean data, but learning from noisy labels significantly degrades performances and remains challenging. Unlike previous works constrained by many conditions, making them infeasible to real noisy cases, this work presents a novel deep self-learning framework to train a robust network on the real noisy datasets without extra supervision. The proposed approach has several appealing benefits. (1) Different from most existing work, it does not rely on any assumption on the distribution of the noisy labels, making it robust to real noises. (2) It does not need extra clean supervision or accessorial network to help training. (3) A self-learning framework is proposed to train the network in an iterative end-to-end manner, which is effective and efficient. Extensive experiments in challenging benchmarks such as Clothing1M and Food101-N show that our approach outperforms its counterparts in all empirical settings.

A great number of invaluable historical photographs unfortunately only exist in the form of halftone prints in old publications such as newspapers or books. Their original continuous-tone films have long been lost or irreparably damaged. There have been attempts to digitally restore these vintage halftone prints to the original film quality or higher. However, even using powerful deep convolutional neural networks, it is still difficult to obtain satisfactory results. The main challenge is that the degradation process is complex and compounded while little to no real data is available for properly training a data-driven method. In this research, we adopt a novel strategy of two-stage deep learning, in which the restoration task is divided into two stages: the removal of printing artifacts and the inverse of halftoning. The advantage of our technique is that only the simple first stage requires unsupervised training in order to make the combined network generalize on real halftone prints, while the more complex second stage of inverse halftoning can be easily trained with synthetic data. Extensive experimental results demonstrate the efficacy of the proposed technique for real halftone prints; the new technique significantly outperforms the existing ones in visual quality.

We introduce a general method of performing Residual Network inference and learning in the JPEG transform domain that allows the network to consume compressed images as input. Our formulation leverages the linearity of the JPEG transform to redefine convolution and batch normalization with a tune-able numerical approximation for ReLu. The result is mathematically equivalent to the spatial domain network up to the ReLu approximation accuracy. A formulation for image classification and a model conversion algorithm for spatial domain networks are given as examples of the method. We show that the sparsity of the JPEG format allows for faster processing of images with little to no penalty in the network accuracy.

We present a method to estimate lighting from a single image of an indoor scene. Previous work has used an environment map representation that does not account for the localized nature of indoor lighting. Instead, we represent lighting as a set of discrete 3D lights with geometric and photometric parameters. We train a deep neural network to regress these parameters from a single image, on a dataset of environment maps annotated with depth. We propose a differentiable layer to convert these parameters to an environment map to compute our loss; this bypasses the challenge of establishing correspondences between estimated and ground truth lights. We demonstrate, via quantitative and qualitative evaluations, that our representation and training scheme lead to more accurate results compared to previous work, while allowing for more realistic 3D object compositing with spatially-varying lighting.

Depth estimation and 3D object detection are critical for scene understanding but remain challenging to perform with a single image due to the loss of 3D information during image capture. Recent models using deep neural networks have improved monocular depth estimation performance, but there is still difficulty in predicting absolute depth and generalizing outside a standard dataset. Here we introduce the paradigm of deep optics, i.e. end-to-end design of optics and image processing, to the monocular depth estimation problem, using coded defocus blur as an additional depth cue to be decoded by a neural network. We evaluate several optical coding strategies along with an end-to-end optimization scheme for depth estimation on three datasets, including NYU Depth v2 and KITTI. We find an optimized freeform lens design yields the best results, but chromatic aberration from a singlet lens offers significantly improved performance as well. We build a physical prototype and validate that chromatic aberrations improve depth estimation on real-world results. In addition, we train object detection networks on the KITTI dataset and show that the lens optimized for depth estimation also results in improved 3D object detection performance.

Texture recognition is a challenging visual task as multiple perceptual attributes may be perceived from the same texture image when combined with different spatial context. Some recent works building upon Convolutional Neural Network (CNN) incorporate feature encoding with orderless aggregating to provide invariance to spatial layouts. However, these existing methods ignore visual texture attributes, which are important cues for describing the real-world texture images, resulting in incomplete description and inaccurate recognition. To address this problem, we propose a novel deep Multiple-Attribute-Perceived Network (MAP-Net) by progressively learning visual texture attributes in a mutually reinforced manner. Specifically, a multi-branch network architecture is devised, in which cascaded global contexts are learned by introducing similarity constraint at each branch, and leveraged as guidance of spatial feature encoding at next branch through an attribute transfer scheme. To enhance the modeling capability of spatial transformation, a deformable pooling strategy is introduced to augment the spatial sampling with adaptive offsets to the global context, leading to perceive new visual attributes. An attribute fusion module is then introduced to jointly utilize the perceived visual attributes and the abstracted semantic concepts at each branch. Experimental results on the five most challenging texture recognition datasets have demonstrated the superiority of the proposed model against the state-of-the-arts.

This paper presents a deep multi-model fusion network to attentively integrate multiple models to separate layers and boost the performance in single-image dehazing. To do so, we first formulate the attentional feature integration module to maximize the integration of the convolutional neural network (CNN) features at different CNN layers and generate the attentional multi-level integrated features (AMLIF). Then, from the AMLIF, we further predict a haze-free result for an atmospheric scattering model, as well as for four haze-layer separation models, and then fuse the results together to produce the final haze-free image. To evaluate the effectiveness of our method, we compare our network with several state-of-the-art methods on two widely-used dehazing benchmark datasets, as well as on two sets of real-world hazy images. Experimental results demonstrate clear quantitative and qualitative improvements of our method over the state-of-the-arts.

Deep metric learning, in which the loss function plays a key role, has proven to be extremely useful in visual recognition tasks. However, existing deep metric learning loss functions such as contrastive loss and triplet loss usually rely on delicately selected samples (pairs or triplets) for fast convergence. In this paper, we propose a new deep metric learning loss function, tuplet margin loss, using randomly selected samples from each mini-batch. Specifically, the proposed tuplet margin loss implicitly up-weights hard samples and down-weights easy samples, while a slack margin in angular space is introduced to mitigate the problem of overfitting on the hardest sample. Furthermore, we address the problem of intra-pair variation by disentangling class-specific information to improve the generalizability of tuplet margin loss. Experimental results on three widely used deep metric learning datasets, CARS196, CUB200-2011, and Stanford Online Products, demonstrate significant improvements over existing deep metric learning methods.

In this paper, we present a deep meta metric learning (DMML) approach for visual recognition. Unlike most existing deep metric learning methods formulating the learning process by an overall objective, our DMML formulates the metric learning in a meta way, and proves that softmax and triplet loss are consistent in the meta space. Specifically, we sample some subsets from the original training set and learn metrics across different subsets. In each sampled sub-task, we split the training data into a support set as well as a query set, and learn the set-based distance, instead of sample-based one, to verify the query cell from multiple support cells. In addition, we introduce hard sample mining for set-based distance to encourage the intra-class compactness. Experimental results on three visual recognition applications including person re-identification, vehicle re-identification and face verification show that the proposed DMML method outperforms most existing approaches.

In this paper, we propose a novel on-line visual tracking framework based on the Siamese matching network and meta-learner network, which run at real-time speeds. Conventional deep convolutional feature-based discriminative visual tracking algorithms require continuous re-training of classifiers or correlation filters, which involve solving complex optimization tasks to adapt to the new appearance of a target object. To alleviate this complex process, our proposed algorithm incorporates and utilizes a meta-learner network to provide the matching network with new appearance information of the target objects by adding target-aware feature space. The parameters for the target-specific feature space are provided instantly from a single forward-pass of the meta-learner network. By eliminating the necessity of continuously solving complex optimization tasks in the course of tracking, experimental results demonstrate that our algorithm performs at a real-time speed while maintaining competitive performance among other state-of-the-art tracking algorithms.

We present a new method for 3D shape reconstruction from a single image, in which a deep neural network directly maps an image to a vector of network weights. The network parametrized by these weights represents a 3D shape by classifying every point in the volume as either within or outside the shape. The new representation has virtually unlimited capacity and resolution, and can have an arbitrary topology. Our experiments show that it leads to more accurate shape inference from a 2D projection than the existing methods, including voxel-, silhouette-, and mesh-based methods. The code will be available at: https: //github.com/gidilittwin/Deep-Meta.

Reconstructing the 3D mesh of a general object from a single image is now possible thanks to the latest advances of deep learning technologies. However, due to the nontrivial difficulty of generating a feasible mesh structure, the state-of-the-art approaches often simplify the problem by learning the displacements of a template mesh that deforms it to the target surface. Though reconstructing a 3D shape with complex topology can be achieved by deforming multiple mesh patches, it remains difficult to stitch the results to ensure a high meshing quality. In this paper, we present an end-to-end single-view mesh reconstruction framework that is able to generate high-quality meshes with complex topologies from a single genus-0 template mesh. The key to our approach is a novel progressive shaping framework that alternates between mesh deformation and topology modification. While a deformation network predicts the per-vertex translations that reduce the gap between the reconstructed mesh and the ground truth, a novel topology modification network is employed to prune the error-prone faces, enabling the evolution of topology. By iterating over the two procedures, one can progressively modify the mesh topology while achieving higher reconstruction accuracy. Moreover, a boundary refinement network is designed to refine the boundary conditions to further improve the visual quality of the reconstructed mesh. Extensive experiments demonstrate that our approach outperforms the current state-of-the-art methods both qualitatively and quantitatively, especially for the shapes with complex topologies.