ECCV 2018 论文列表

Semantic segmentation of outdoor scenes is problematic when there are variations in imaging conditions. It is known that albedo (reflectance) is invariant to all kinds of illumination effects. Thus, using reflectance images for semantic segmentation task can be favorable. Additionally, not only segmentation may benefit from reflectance, but also segmentation may be useful for reflectance computation. Therefore, in this paper, the tasks of semantic segmentation and intrinsic image decomposition are considered as a combined process by exploring their mutual relationship in a joint fashion. To that end, we propose a supervised end-to-end CNN architecture to jointly learn intrinsic image decomposition and semantic segmentation. We analyze the gains of addressing those two problems jointly. Moreover, new cascade CNN architectures for intrinsic-for-segmentation and segmentation-for-intrinsic are proposed as single tasks. Furthermore, a dataset of 35K synthetic images of natural environments is created with corresponding albedo and shading (intrinsics), as well as semantic labels (segmentation) assigned to each object/scene. The experiments show that joint learning of intrinsic image decomposition and semantic segmentation is beneficial for both tasks for natural scenes. Dataset and models are available at: https://ivi.fnwi.uva.nl/cv/intrinseg

Hyperspectral image (HSI) recovery from a single RGB image has attracted much attention, whose performance has recently been shown to be sensitive to the camera spectral sensitivity (CSS). In this paper, we present an efficient convolutional neural network (CNN) based method, which can jointly select the optimal CSS from a candidate dataset and learn a mapping to recover HSI from a single RGB image captured with this algorithmically selected camera. Given a specific CSS, we first present a HSI recovery network, which accounts for the underlying characteristics of the HSI, including spectral nonlinear mapping and spatial similarity. Later, we append a CSS selection layer onto the recovery network, and the optimal CSS can thus be automatically determined from the network weights under the nonnegative sparse constraint. Experimental results show that our HSI recovery network outperforms state-of-the-art methods in terms of both quantitative metrics and perceptive quality, and the selection layer always returns a CSS consistent to the best one determined by exhaustive search.

Removing camera motion blur from a single light field is a challenging task since it is highly ill-posed inverse problem. The problem becomes even worse when blur kernel varies spatially due to scene depth variation and high-order camera motion. In this paper, we propose a novel algorithm to estimate all blur model variables jointly, including latent sub-aperture image, camera motion, and scene depth from the blurred 4D light field. Exploiting multi-view nature of a light field relieves the inverse property of the optimization by utilizing strong depth cues and multi-view blur observation. The proposed joint estimation achieves high quality light field deblurring and depth estimation simultaneously under arbitrary 6-DOF camera motion and unconstrained scene depth. Intensive experiment on real and synthetic blurred light field confirms that the proposed algorithm outperforms the state-of-the-art light field deblurring and depth estimation methods.

We present a novel method that is able to track a complex deformable object in interaction with a hand. This is achieved by formulating and solving an optimization problem that jointly considers the hand, the deformable object and the hand/object contact points. The optimization evaluates several hand/object contact configuration hypotheses and adopts the one that results in the best fit of the object's model to the available RGBD observations in the vicinity of the hand. Thus, the hand is not treated as a distractor that occludes parts of the deformable object, but as a source of valuable information. Experimental results on a dataset that has been developed specifically for this new problem illustrate the superior performance of the proposed approach against relevant, state of the art solutions.

We propose a straightforward method that simultaneously reconstructs the 3D facial structure and provides dense alignment. To achieve this, we design a 2D representation called UV position map which records the 3D shape of a complete face in UV space, then train a simple Convolutional Neural Network to regress it from a single 2D image. We also integrate a weight mask into the loss function during training to improve the performance of the network. Our method does not rely on any prior face model, and can reconstruct full facial geometry along with semantic meaning. Meanwhile, our network is very light-weighted and spends only 9.8ms to process an image, which is extremely faster than previous works. Experiments on multiple challenging datasets show that our method surpasses other state-of-the-art methods on both reconstruction and alignment tasks by a large margin. Code is available at https://github.com/YadiraF/PRNet.

Despite the fact that nonlinear subspace learning techniques (e.g. manifold learning) have successfully applied to data representation, there is still room for improvement in explainability (explicit mapping), generalization (out-of-samples), and cost-effectiveness (linearization). To this end, a novel linearized subspace learning technique is developed in a joint and progressive way, called joint and progressive learning strategy (J-Play), with its application to multi-label classification. The J-Play learns high-level and semantically meaningful feature representation from high-dimensional data by 1) jointly performing multiple subspace learning and classification to find a latent subspace where samples are expected to be better classified; 2) progressively learning multi-coupled projections to linearly approach the optimal mapping bridging the original space with the most discriminative subspace; 3) locally embedding manifold structure in each learnable latent subspace. Extensive experiments are performed to demonstrate the superiority and effectiveness of the proposed method in comparison with previous state-of-the-art methods.

In this work, we tackle the problem of crowd counting in images. We present a Convolutional Neural Network (CNN) based density estimation approach to solve this problem. Predicting a high resolution density map in one go is a challenging task. Hence, we present a two branch CNN architecture for generating high resolution density maps, where the first branch generates a low resolution density map and the second branch incorporates the prediction and feature maps from the first branch to generate a high resolution density map. We also propose a multi-stage extension of our approach where each stage in the pipeline utilizes the predictions from all the previous stages. Empirical comparison with the previous state-of-the-art crowd counting methods shows that our method achieves the lowest mean absolute error on three challenging crowd counting benchmarks: Shanghaitech, WorldExpo'10, and UCF datasets.

The prediction accuracy has been the long-lasting and sole standard for comparing the performance of different image classification models, including the ImageNet competition. However, recent studies have highlighted the lack of robustness in well-trained deep neural networks to adversarial examples. Visually imperceptible perturbations to natural images can easily be crafted and mislead the image classifiers towards misclassification. To demystify the trade-offs between robustness and accuracy, in this paper we thoroughly benchmark 18 ImageNet models using multiple robustness metrics, including the distortion, success rate and transferability of adversarial examples between 306 pairs of models. Our extensive experimental results reveal several new insights: (1) linear scaling law - the empirical $ell_2$ and $ell_infty$ distortion metrics scale linearly with the $log$ value of classification error; (2) model architecture is a more critical factor to robustness than model size, and the disclosed accuracy-robustness Pareto frontier can be used as an evaluation criterion for ImageNet model designers; (3) for a similar network architecture, increasing network depth slightly improves robustness in $ell_infty$ distortion; (4) there exist models (in VGG family) that exhibit high adversarial transferability, while most adversarial examples crafted from one model can only be transferred within the same family. Experiment code is publicly available at url{https://github.com/huanzhang12/Adversarial_Survey}.

We present a joint Structure-Stereo optimization model that is robust for disparity estimation under low-light conditions. Eschewing the traditional denoising approach - which we show to be ineffective for stereo due to its artefacts and the questionable use of the PSNR metric, we propose to instead rely on structures comprising of piecewise constant regions and principal edges in the given image, as these are the important regions for extracting disparity information. We also judiciously retain the coarser textures for stereo matching, discarding the finer textures as they are apt to be inextricably mixed with noise. This selection process in the structure-texture decomposition step is aided by the stereo matching constraint in our joint Structure-Stereo formulation. The resulting optimization problem is complex but we are able to decompose it into sub-problems that admit relatively standard solutions. Our experiments confirm that our joint model significantly outperforms the baseline methods on both synthetic and real noise datasets.

Humans have a remarkable ability to use physical commonsense and predict the effect of collisions. But do they understand the underlying factors? Can they predict if the underlying factors have changed? Interestingly, in most cases humans can predict the effects of similar collisions with different conditions such as changes in mass, friction, etc. It is postulated this is primarily because we learn to model physics with meaningful latent variables. This does not imply we can estimate the precise values of these meaningful variables (estimate exact values of mass or friction). Inspired by this observation, we propose an interpretable intuitive physics model where specific dimensions in the bottleneck layers correspond to different physical properties. In order to demonstrate that our system models these underlying physical properties, we train our model on collisions of different shapes (cube, cone, cylinder, spheres etc.) and test on collisions of unseen combinations of shapes. Furthermore, we demonstrate our model generalizes well even when similar scenes are simulated with different underlying properties.

Explanations of the decisions made by a deep neural network are important for human end-users to be able to understand and diagnose the trustworthiness of the system. Current neural networks used for visual recognition are generally used as black boxes that do not provide any human interpretable justification for a prediction. In this work we propose a new framework called Interpretable Basis Decomposition for providing visual explanations for classification networks. By decomposing the neural activations of the input image into semantically interpretable components pre-trained from a large concept corpus, the proposed framework is able to disentangle the evidence encoded in the activation feature vector, and quantify the contribution of each piece of evidence to the final prediction. We apply our framework for providing explanations to several popular networks for visual recognition, and show it is able to explain the predictions given by the networks in human-interpretable way. The human interpretability of the visual explanations provided by our framework and other recent explanation methods is evaluated through Amazon Mechanical Turk, showing that our framework generates more faithful and interpretable explanations.

Perceiving a visual concept as a mixture of learned ones is natural for humans, aiding them to grasp new concepts and strengthening old ones. For all their power and recent success, deep convolutional networks do not have this ability. Inspired by recent work on universal representations for neural networks, we propose a simple emulation of this mechanism by purposing batch normalization layers to discriminate visual classes, and formulating a way to combine them to solve new tasks. We show that this can be applied for 2-way few-shot learning where we obtain between 4% and 17% better accuracy compared to straightforward full fine-tuning, and demonstrate that it can also be extended to the orthogonal application of style transfer.

Interactive image segmentation is critical for many image editing tasks. While recent advanced methods on interactive segmentation focus on the region-based paradigm, more traditional boundary-based methods such as Intelligent Scissor are still popular in practice as they allow users to have active control of the object boundaries. Existing methods for boundary-based segmentation solely rely on low-level image features, such as edges for boundary extraction, which limits their ability to adapt to high-level image content and user intention. In this paper, we introduce an interaction-aware method for boundary-based image segmentation. Instead of relying on pre-defined low-level image features, our method adaptively predicts object boundaries according to image content and user interactions. Therein, we develop a fully convolutional encoder-decoder network that takes both the image and user interactions (e.g. clicks on boundary points) as input and predicts semantically meaningful boundaries that match user intentions. Our method explicitly models the dependency of boundary extraction results on image content and user interactions. Experiments on two public interactive segmentation benchmarks show that our method significantly improves the boundary quality of segmentation results compared to state-of-the-art methods while requiring fewer user interactions.

Local features at neighboring spatial positions in feature maps have high correlation since their receptive fields are often overlapped. Self-attention usually uses the weighted sum (or other functions) with internal elements of each local feature to obtain its weight score, which ignores interactions among local features. To address this, we propose an effective interaction-aware self-attention model inspired by PCA to learn attention maps. Furthermore, since different layers in a deep network capture feature maps of different scales, we use these feature maps to construct a spatial pyramid and then utilize multi-scale information to obtain more accurate attention scores, which are used to weight the local features in all spatial positions of feature maps to calculate attention maps. Moreover, our spatial pyramid attention is unrestricted to the number of its input feature maps so it is easily extended to a spatio-temporal version. Finally, our model is embedded in general CNNs to form end-to-end attention networks for action classification. Experimental results show that our method achieves the state-of-the-art results on the UCF101, HMDB51 and untrimmed Charades.

Videos recorded from first person (egocentric) perspective have little visual appearance in common with those from third person perspective, especially with videos captured by top-view surveillance cameras. In this paper, we aim to relate these two sources of information from a surveillance standpoint, namely in terms of identification and temporal alignment. Given an egocentric video and a top-view video, our goals are to: a) identify the egocentric camera holder in the top-view video (self-identification), b) identify the humans visible in the content of the egocentric video, within the content of the top-view video (re-identification), and c) temporally align the two videos. The main challenge is that each of these tasks is highly dependent on the other two. We propose a unified framework to jointly solve all three problems. We evaluate the efficacy of the proposed approach on a publicly available dataset containing a variety of videos recorded in different scenarios.

State-of-the-art human pose estimation methods are based on heat map representation. In spite of the good performance, the representation has a few issues in nature, such as non-differentiable post-processing and quantization error. This work shows that a simple integral operation relates and unifies the heat map representation and joint regression, thus avoiding the above issues. It is differentiable, efficient, and compatible with any heat map based methods. Its effectiveness is convincingly validated via comprehensive ablation experiments under various settings, specifically on 3D pose estimation, for the first time.

Instance-level human parsing towards real-world human analysis scenarios is still under-explored due to the absence of sufficient data resources and technical difficulty in parsing multiple instances in a single pass. Several related works all follow the ``parsing-by-detection" pipeline that heavily relies on separately trained detection models to localize instances and then performs human parsing for each instance sequentially. Nonetheless, two discrepant optimization targets of detection and parsing lead to suboptimal representation learning and error accumulation for final results. In this work, we make the first attempt to explore a detection-free Part Grouping Network (PGN) for efficiently parsing multiple people in an image in a single pass. Our PGN reformulates instance-level human parsing as two twinned sub-tasks that can be jointly learned and mutually refined via a unified network: 1) semantic part segmentation for assigning each pixel as a human part (e.g., face, arms); 2) instance-aware edge detection to group semantic parts into distinct person instances. Thus the shared intermediate representation would be endowed with capabilities in both characterizing fine-grained parts and inferring instance belongings of each part. Finally, a simple instance partition process is employed to get final results during inference. We conducted experiments on PASCAL-Person-Part dataset and our PGN outperforms all state-of-the-art methods. Furthermore, we show its superiority on a newly collected multi-person parsing dataset (CIHP) including 38,280 diverse images, which is the largest dataset so far and can facilitate more advanced human analysis.

Image-based generative methods, such as generative adversarial networks (GANs) have already been able to generate realistic images with much context control, specially when they are conditioned. However, most successful frameworks share a common procedure which performs an image-to-image translation with pose of figures in the image untouched. When the objective is reposing a figure in an image while preserving the rest of the image, the state-of-the-art mainly assumes a single rigid body with simple background and limited pose shift, which can hardly be extended to the images under normal settings. In this paper, we introduce an image ``inner space'' preserving model that assigns an interpretable low-dimensional pose descriptor (LDPD) to an articulated figure in the image. Figure reposing is then generated by passing the LDPD and the original image through multi-stage augmented hourglass networks in a conditional GAN structure, called inner space preserving generative pose machine (ISP-GPM). We evaluated ISP-GPM on reposing human figures, which are highly articulated with versatile variations. Test of a state-of-the-art pose estimator on our reposed dataset gave an accuracy over 80% on PCK0.5 metric. The results also elucidated that our ISP-GPM is able to preserve the background with high accuracy while reasonably recovering the area blocked by the figure to be reposed.

The perspective camera and the isometric surface prior have recently gathered increased attention for Non-Rigid Structure-from-Motion (NRSfM). De- spite the recent progress, several challenges remain, particularly the computa- tional complexity and the unknown camera focal length. In this paper we present a method for incremental Non-Rigid Structure-from-Motion (NRSfM) with the perspective camera model and the isometric surface prior with unknown focal length. In the template-based case, we provide a method to estimate four param- eters of the camera intrinsics. For the template-less scenario of NRSfM, we pro- pose a method to upgrade reconstructions obtained for one focal length to another based on local rigidity and the so-called Maximum Depth Heuristics (MDH). On its basis we propose a method to simultaneously recover the focal length and the non-rigid shapes. We further solve the problem of incorporating a large number of points and adding more views in MDH-based NRSfM and efficiently solve them with Second-Order Cone Programming (SOCP). This does not require any shape initialization and produces results orders of times faster than many methods. We provide evaluations on standard sequences with ground-truth and qualitative re- constructions on challenging YouTube videos. These evaluations show that our method performs better in both speed and accuracy than the state of the art.