CVPR 2015 论文列表

We present a method for simulating makeup in a face image. To generate realistic results without detailed geometric and reflectance measurements of the user, we propose to separate the image into intrinsic image layers and alter them according to proposed adaptations of physically-based reflectance models. Through this layer manipulation, the measured properties of cosmetic products are applied while preserving the appearance characteristics and lighting conditions of the target face. This approach is demonstrated on various forms of cosmetics including foundation, blush, lipstick, and eye shadow. Experimental results exhibit a close approximation to ground truth images, without artifacts such as transferred personal features and lighting effects that degrade the results of image-based makeup transfer methods.

We propose a novel approach to compute the camera pose with respect to a reference object given only mirrored views. The latter originate from a planar mirror at different unknown poses. This problem is highly relevant in several extrinsic camera calibration scenarios, where the camera cannot see the reference object directly. In contrast to numerous existing methods, our approach does not employ the fixed axis rotation constraint, but represents a more elegant formulation as a rotation averaging problem. Our theoretical contribution extends the applicability of rotation averaging to a more general case, and enables mirror-based pose estimation in closed-form under the chordal L2-metric, or in an outlier-robust way by employing iterative L1-norm averaging. We demonstrate the advantages of our approach on both synthetic and real data, and show how the method can be applied to calibrate the non-overlapping pair of cameras of a common smart phone.

In this paper, we address the person re-identification problem, discovering the correct matches for a probe person image from a set of gallery person images. We follow the learning-to-rank methodology and learn a similarity function to maximize the difference between the similarity scores of matched and unmatched images for a same person. We introduce at least three contributions to person re-identification. First, we present an explicit polynomial kernel feature map, which is capable of characterizing the similarity information of all pairs of patches between two images, called soft-patch-match, instead of greedily keeping only the best matched patch, and thus more robust. Second, we introduce a mixture of linear similarity functions that is able to discover different soft-patch-matching patterns. Last, we introduce a negative semi-definite regularization over a subset of the weights in the similarity function, which is motivated by the connection between explicit polynomial kernel feature map and the Mahalanobis distance, as well as the sparsity constraint over the parameters to avoid over-fitting. Experimental results over three public benchmarks demonstrate the superiority of our approach.

Automatically describing the content of an image is a fundamental problem in artificial intelligence that connects computer vision and natural language processing. In this paper, we present a generative model based on a deep recurrent architecture that combines recent advances in computer vision and machine translation and that can be used to generate natural sentences describing an image. The model is trained to maximize the likelihood of the target description sentence given the training image. Experiments on several datasets show the accuracy of the model and the fluency of the language it learns solely from image descriptions. Our model is often quite accurate, which we verify both qualitatively and quantitatively. For instance, while the current state-of-the-art BLEU score (the higher the better) on the Pascal dataset is 25, our approach yields 59, to be compared to human performance around 69. We also show BLEU score improvements on Flickr30k, from 56 to 66, and on SBU, from 19 to 28. Lastly, on the newly released COCO dataset, we achieve a BLEU-4 of 27.7, which is the current state-of-the-art.

We propose new dense descriptors for texture segmentation. Given a region of arbitrary shape in an image, these descriptors are formed from shape-dependent scale spaces of oriented gradients. These scale spaces are defined by Poisson-like partial differential equations. A key property of our new descriptors is that they do not aggregate image data across the boundary of the region, in contrast to existing descriptors based on aggregation of oriented gradients. As an example, we show how the descriptor can be incorporated in a Mumford-Shah energy for texture segmentation. We test our method on several challenging datasets for texture segmentation and textured object tracking. Experiments indicate that our descriptors lead to more accurate segmentation than non-shape dependent descriptors and the state-of-the-art in texture segmentation.

Shape-from-Template (SfT) is the problem of inferring the shape of a deformable object as observed in an image using a shape template. We call 2DSfT the 'usual' instance of SfT where the shape is a surface embedded in 3D and the image a 2D projection. We introduce 1DSfT, a novel instance of SfT where the shape is a curve embedded in 2D and the image a 1D projection. We focus on isometric deformations, for which 2DSfT is a well-posed problem, and admits an analytical local solution which may be used to initialize nonconvex refinement. Through a complete theoretical study of 1DSfT with perspective projection, we show that it is related to 2DSfT, but may have very different properties: (i) 1DSfT cannot be exactly solved locally and (ii) 1DSfT cannot be solved uniquely, as it has a discrete amount of at least two solutions. We then propose two convex initialization algorithms, a local analytical one based on infinitesimal planarity and a global one based on inextensibility. We show how nonconvex refinement can be implemented where, contrarily to current 2DSfT methods, one may enforce isometry exactly using a novel angle-based parameterization. Finally, our method is tested with simulated and real data.

We present an algorithm for automatic detection of a large number of anthropometric landmarks on 3D faces. Our approach does not use texture and is completely shape based in order to detect landmarks that are morphologically significant. The proposed algorithm evolves level set curves with adaptive geometric speed functions to automatically extract effective seed points for dense correspondence. Correspondences are established by minimizing the bending energy between patches around seed points of given faces to those of a reference face. Given its hierarchical structure, our algorithm is capable of establishing thousands of correspondences between a large number of faces. Finally, a morphable model based on the dense corresponding points is fitted to an unseen query face for transfer of correspondences and hence automatic detection of landmarks. The proposed algorithm can detect any number of pre-defined landmarks including subtle landmarks that are even difficult to detect manually. Extensive experimental comparison on two benchmark databases containing 6,507 scans shows that our algorithm outperforms six state of the art algorithms.

We introduce a method to recover the shape of a smooth dielectric object from polarization images taken with a light source from different directions. We present two constraints on shading and polarization and use both in a single optimization scheme. This integration is motivated by the fact that photometric stereo and polarization-based methods have complementary abilities. The polarization-based method can give strong cues for the surface orientation and refractive index, which are independent of the light direction. However, it has ambiguities in selecting between two ambiguous choices of the surface orientation, in the relationship between refractive index and zenith angle (observing angle), and limited performance for surface points with small zenith angles, where the polarization effect is weak. In contrast, photometric stereo method with multiple light sources can disambiguate the surface orientation and give a strong relationship between the surface normals and light directions. However, it has limited performance for large zenith angles, refractive index estimation, and faces the ambiguity in case the light direction is unknown. Taking their advantages, our proposed method can recover the surface normals for both small and large zenith angles, the light directions, and the refractive indexes of the object. The proposed method is successfully evaluated by simulation and real-world experiments.

One key challenge of facial traits recognition is the large non-rigid appearance variations due to irrelevant real world factors, such as viewpoint and expression changes. In this paper, we explore how the shape information, i.e. facial landmark positions, can be explicitly deployed into the popular Convolutional Neural Network (CNN) architecture to disentangling such irrelevant non-rigid appearance variations. First, instead of using fixed kernels, we propose a kernel adaptation method to dynamically determine the convolutional kernels according to the distribution of facial landmarks, which helps learning more robust features. Second, motivated by the intuition that different local facial regions may demand different adaptation functions, we further propose a tree-structured convolutional architecture to hierarchically fuse multiple local adaptive CNN subnetworks. Comprehensive experiments on WebFace, Morph II and MultiPIE databases well validate the effectiveness of the proposed kernel adaptation method and tree-structured convolutional architecture for facial traits recognition tasks, including identity, age and gender classification. For all the tasks, the proposed architecture consistently achieves the state-of-the-art performances.

We present a novel learning-based framework for shadow detection from a single image. The local structure of shadow boundaries as well as the global interactions of the shadow and non-shadow regions remain largely unexploited by previous learning-based approaches. In this paper, we propose an efficient structured labelling framework for shadow detection from a single image. A convolutional Neural Networks framework is designed to capture the local structure information of shadow edge and to learn the most relevant features. We further propose and formulate a global shadow optimization framework which can model the complex global interactions over the shadow and light regions. Using the shadow edges detected by our proposed method, the shadow map can be solved by efficient least-square optimization. Our proposed framework is efficient and achieves state-of-the-art results on the major shadow benchmark databases collected under a variety of conditions.

Objects' spatial layout estimation and clutter identification are two important tasks to understand indoor scenes. We propose to solve both of these problems in a joint framework using RGBD images of indoor scenes. In contrast to recent approaches which focus on either one of these two problems, we perform `fine grained structure categorization' by predicting all the major objects and simultaneously labeling the cluttered regions. A conditional random field model is proposed to incorporate a rich set of local appearance, geometric features and interactions between the scene elements. We take a structural learning approach with a loss of 3D localisation to estimate the model parameters from a large annotated RGBD dataset, and a mixed integer linear programming formulation for inference. We demonstrate that our approach is able to detect cuboids and estimate cluttered regions across many different object and scene categories in the presence of occlusion, illumination and appearance variations.

We present a co-clustering framework that can be used to discover multiple semantic and visual senses of a given Noun Phrase (NP). Unlike traditional clustering approaches which assume a one-to-one mapping between the clusters in the text-based feature space and the visual space, we adopt a one-to-many mapping between the two spaces. This is primarily because each semantic sense (concept) can correspond to different visual senses due to viewpoint and appearance variations. Our structure-EM style optimization not only extracts the multiple senses in both semantic and visual feature space, but also discovers the mapping between the senses. We introduce a challenging dataset (CMU Polysemy-30) for this problem consisting of 30 NPs ($\sim$5600 labeled instances out of $\sim$22K total instances). We have also conducted a large-scale experiment that performs sense disambiguation for $\sim$2000 NPs.

Multi-label problems arise in various domains including automatic multimedia data categorization, and have generated significant interest in computer vision and machine learning community. However, existing methods do not adequately address two key challenges: exploiting correlations between labels and making up for the lack of labeled data or even missing labels. In this paper, we proposed a semi-supervised low-rank mapping (SLRM) model to handle these two challenges. SLRM model takes advantage of the nuclear norm regularization on mapping to effectively capture the label correlations. Meanwhile, it introduces manifold regularizer on mapping to capture the intrinsic structure among data, which provides a good way to reduce the required labeled data with improving the classification performance. Furthermore, we designed an efficient algorithm to solve SLRM model based on alternating direction method of multipliers and thus it can efficiently deal with large-scale data sets. Experiments on four real-world multimedia data sets demonstrate that the proposed method can exploit the label correlations and obtain promising and better label prediction results than state-of-the-art methods.

In many learning tasks, the structure of the target space of a function holds rich information about the relationships between evaluations of functions on different data points. Existing approaches attempt to exploit this relationship information implicitly by enforcing smoothness on function evaluations only. However, what happens if we explicitly regularize the relationships between function evaluations? Inspired by homophily, we regularize based on a smooth relationship function, either defined from the data or with labels. In experiments, we demonstrate that this significantly improves the performance of state-of-the-art algorithms in semi-supervised classification and in spectral data embedding for constrained clustering and dimensionality reduction.

In many real-world applications, we are often facing the problem of cross domain learning, i.e., to borrow the labeled data or transfer the already learnt knowledge from a source domain to a target domain. However, simply applying existing source data or knowledge may even hurt the performance, especially when the data distribution in the source and target domain is quite different, or there are very few labeled data available in the target domain. This paper proposes a novel domain adaptation framework, named Semi-supervised Domain Adaptation with Subspace Learning (SDASL), which jointly explores invariant low-dimensional structures across domains to correct data distribution mismatch and leverages available unlabeled target examples to exploit the underlying intrinsic information in the target domain. Specifically, SDASL conducts the learning by simultaneously minimizing the classification error, preserving the structure within and across domains, and restricting similarity defined on unlabeled target examples. Encouraging results are reported for two challenging domain transfer tasks (including image-to-image and image-to-video transfers) on several standard datasets in the context of both image object recognition and video concept detection.

With benefits of low storage costs and high query speeds, hashing methods are widely researched for efficiently retrieving large-scale data, which commonly contains multiple views, e.g. a news report with images, videos and texts. In this paper, we study the problem of cross-view retrieval and propose an effective Semantics-Preserving Hashing method, termed SePH. Given semantic affinities of training data as supervised information, SePH transforms them into a probability distribution and approximates it with to-be-learnt hash codes in Hamming space via minimizing the Kullback-Leibler divergence. Then kernel logistic regression with a sampling strategy is utilized to learn the nonlinear projections from features in each view to the learnt hash codes. And for any unseen instance, predicted hash codes and their corresponding output probabilities from observed views are utilized to determine its unified hash code, using a novel probabilistic approach. Extensive experiments conducted on three benchmark datasets well demonstrate the effectiveness and reasonableness of SePH.

In this paper, we study the problem of semantic part segmentation for animals. This is more challenging than standard object detection, object segmentation and pose estimation tasks because semantic parts of animals often have similar appearance and highly varying shapes. To tackle these challenges, we build a mixture of compositional models to represent the object boundary and the boundaries of semantic parts. And we incorporate edge, appearance, and semantic part cues into the compositional model. Given part-level segmentation annotation, we develop a novel algorithm to learn a mixture of compositional models under various poses and viewpoints for certain animal classes. Furthermore, a linear complexity algorithm is offered for efficient inference of the compositional model using dynamic programming. We evaluate our method for horse and cow using a newly annotated dataset on Pascal VOC 2010 which has pixelwise part labels. Experimental results demonstrate the effectiveness of our method.

Semantic object segmentation in video is an important step for large-scale multimedia analysis. In many cases, however, semantic objects are only tagged at video-level, making them difficult to be located and segmented. To address this problem, this paper proposes an approach to segment semantic objects in weakly labeled video via object detection. In our approach, a novel video segmentationby-detection framework is proposed, which first incorporates object and region detectors pre-trained on still images to generate a set of detection and segmentation proposals. Based on the noisy proposals, several object tracks are then initialized by solving a joint binary optimization problem with min-cost flow. As such tracks actually provide rough configurations of semantic objects, we thus refine the object segmentation while preserving the spatiotemporal consistency by inferring the shape likelihoods of pixels from the statistical information of tracks. Experimental results on Youtube-Objects dataset and SegTrack v2 dataset demonstrate that our method outperforms state-of-the-arts and shows impressive results.

This paper describes an automatic algorithm for global alignment of LiDAR data collected with Google Street View cars in urban environments. The problem is challenging because global pose estimation techniques (GPS) do not work well in city environments with tall buildings, and local tracking techniques (integration of inertial sensors, structure-from-motion, etc.) provide solutions that drift over long ranges, leading to solutions where data collected over wide ranges is warped and misaligned by many meters. Our approach to address this problem is to extract ``semantic features'' with object detectors (e.g., for facades, poles, cars, etc.) that can be matched robustly at different scales, and thus are selected for different iterations of an ICP algorithm. We have implemented an all-to-all, non-rigid, global alignment based on this idea that provides better results than alternatives during experiments with data from large regions of New York, San Francisco, Paris, and Rome.