ICCV 2015 Papers

We present a robust and real-time monocular six degree of freedom relocalization system. Our system trains a convolutional neural network to regress the 6-DOF camera pose from a single RGB image in an end-to-end manner with no need of additional engineering or graph optimisation. The algorithm can operate indoors and outdoors in real time, taking 5ms per frame to compute. It obtains approximately 2m and 3 degrees accuracy for large scale outdoor scenes and 0.5m and 5 degrees accuracy indoors. This is achieved using an efficient 23 layer deep convnet, demonstrating that convnets can be used to solve complicated out of image plane regression problems. This was made possible by leveraging transfer learning from large scale classification data. We show that the PoseNet localizes from high level features and is robust to difficult lighting, motion blur and different camera intrinsics where point based SIFT registration fails. Furthermore we show how the pose feature that is produced generalizes to other scenes allowing us to regress pose with only a few dozen training examples.

Face alignment aims to estimate the locations of a set of landmarks for a given image.This problem has received much attention as evidenced by the recent advancement in both the methodology and performance. However, most of the existing works neither explicitly handle face images with arbitrary poses, nor perform large-scale experiments on non-frontal and profile face images. In order to address these limitations, this paper proposes a novel face alignment algorithm that estimates both 2D and 3D landmarks and their 2D visibilities for a face image with an arbitrary pose. By integrating a 3D point distribution model, a cascaded coupled-regressor approach is designed to estimate both the camera projection matrix and the 3D landmarks. Furthermore, the 3D model also allows us to automatically estimate the 2D landmark visibilities via surface normal. We use a substantially larger collection of all-pose face images to evaluate our algorithm and demonstrate superior performances than the state-of-the-art methods.

We address the task of predicting pose for objects of unannotated object categories from a small seed set of annotated object classes. We present a generalized classifier that can reliably induce pose given a single instance of a novel category. In case of availability of a large collection of novel instances, our approach then jointly reasons over all instances to improve the initial estimates. We empirically validate the various components of our algorithm and quantitatively show that our method produces reliable pose estimates. We also show qualitative results on a diverse set of classes and further demonstrate the applicability of our system for learning shape models of novel object classes.

Multi-camera triangulation of feature points based on a minimisation of the overall L2 reprojection error can get stuck in suboptimal local minima or require slow global optimisation. For this reason, researchers have proposed optimising the L-infinity norm of the L2 single view reprojection errors, which avoids the problem of local minima entirely. In this paper we present a novel method for L-infinity triangulation that minimizes the L-infinity norm of the L-infinity reprojection errors: this apparently small difference leads to a much faster but equally accurate solution which is related to the MLE under the assumption of uniform noise. The proposed method adopts a new optimisation strategy based on solving simple quadratic equations. This stands in contrast with the fastest existing methods, which solve a sequence of more complex auxiliary Linear Programming or Second Order Cone Problems. The proposed algorithm performs well: for triangulation, it achieves the same accuracy as existing techniques while executing faster and being straightforward to implement.

Photometric stereo is widely used for 3D reconstruction. However, its use in scattering media such as water, biological tissue and fog has been limited until now, because of forward scattered light from both the source and object, as well as light scattered back from the medium (backscatter). Here we make three contributions to address the key modes of light propagation, under the common single scattering assumption for dilute media. First, we show through extensive simulations that single-scattered light from a source can be approximated by a point light source with a single direction. This alleviates the need to handle light source blur explicitly. Next, we model the blur due to scattering of light from the object. We measure the object point-spread function and introduce a simple deconvolution method. Finally, we show how imaging fluorescence emission where available, eliminates the backscatter component and increases the signal-to-noise ratio. Experimental results in a water tank, with different concentrations of scattering media added, show that deconvolution produces higher-quality 3D reconstructions than previous techniques, and that when combined with fluorescence, can produce results similar to that in clear water even for highly turbid media.

Most existing successful photometric stereo setups require large angular variations in illumination directions, which results in acquisition rigs that have large spatial extent. For many applications, especially involving mobile devices, it is important that the device be spatially compact. This naturally implies smaller angular variations in the illumination directions. This paper studies the effect of small angular variations in illumination directions to photometric stereo. We explore both theoretical justification and practical issues in the design of a compact and portable photometric stereo device on which a camera is surrounded by a ring of point light sources. We first derive the relationship between the estimation error of surface normal and the baseline of the point light sources. Armed with this theoretical insight, we develop a small baseline photometric stereo prototype to experimentally examine the theory and its practicality.

Photometric stereo (PS) is an established technique for high-detail reconstruction of 3D geometry and appearance. To correct for surface integration errors, PS is often combined with multiview stereo (MVS). With dynamic objects, PS reconstruction also faces the problem of computing optical flow (OF) for image alignment under rapid changes in illumination. Current PS methods typically compute optical flow and MVS as independent stages, each one with its own limitations and errors introduced by early regularization. In contrast, scene flow methods estimate geometry and motion, but lack the fine detail from PS. This paper proposes photogeometric scene flow (PGSF) for high-quality dynamic 3D reconstruction. PGSF performs PS, OF, and MVS simultaneously. It is based on two key observations: (i) while image alignment improves PS, PS allows for surfaces to be relit to improve alignment; (ii) PS provides surface gradients that render the smoothness term in MVS unnecessary, leading to truly data-driven, continuous depth estimates. This synergy is demonstrated in the quality of the resulting RGB appearance, 3D geometry, and 3D motion.

In this paper, we aim to automatically render aging faces in a personalized way. Basically, a set of age-group specific dictionaries are learned, where the dictionary bases corresponding to the same index yet from different dictionaries form a particular aging process pattern cross different age groups, and a linear combination of these patterns expresses a particular personalized aging process. Moreover, two factors are taken into consideration in the dictionary learning process. First, beyond the aging dictionaries, each subject may have extra personalized facial characteristics, e.g. mole, which are invariant in the aging process. Second, it is challenging or even impossible to collect faces of all age groups for a particular subject, yet much easier and more practical to get face pairs from neighboring age groups. Thus a personality-aware coupled reconstruction loss is utilized to learn the dictionaries based on face pairs from neighboring age groups. Extensive experiments well demonstrate the advantages of our proposed solution over other state-of-the-arts in term of personalized aging progression, as well as the performance gain for cross-age face verification by synthesizing aging faces.

Recognising persons in everyday photos presents major challenges (occluded faces, different clothing, locations, etc.) for machine vision. We propose a convnet based person recognition system on which we provide an in-depth analysis of informativeness of different body cues, impact of training data, and the common failure modes of the system. In addition, we discuss the limitations of existing benchmarks and propose more challenging ones. Our method is simple and is built on open source and open data, yet it improves the state of the art results on a large dataset of social media photos (PIPA).

This paper introduces a new approach to address the person re-identification problem in cameras with non-overlapping fields of view. Unlike previous approaches that learn Mahalanobis-like distance metrics in some transformed feature space, we propose to learn a dictionary that is capable of discriminatively and sparsely encoding features representing different people. Our approach directly addresses two key challenges in person re-identification: viewpoint variations and discriminability. First, to tackle viewpoint and associated appearance changes, we learn a single dictionary to represent both gallery and probe images in the training phase. We then discriminatively train the dictionary by enforcing explicit constraints on the associated sparse representations of the feature vectors. In the testing phase, we re-identify a probe image by simply determining the gallery image that has the closest sparse representation to that of the probe image in the Euclidean sense. Extensive performance evaluations on three publicly available multi-shot re-identification datasets demonstrate the advantages of our algorithm over several state-of-the-art dictionary learning, temporal sequence matching, and spatial appearance and metric learning based techniques.

This paper addresses the problem of handling spatial misalignments due to camera-view changes or human-pose variations in person re-identification. We first introduce a boosting-based approach to learn a correspondence structure which indicates the patch-wise matching probabilities between images from a target camera pair. The learned correspondence structure can not only capture the spatial correspondence pattern between cameras but also handle the viewpoint or human-pose variation in individual images. We further introduce a global-based matching process. It integrates a global matching constraint over the learned correspondence structure to exclude cross-view misalignments during the image patch matching process, hence achieving a more reliable matching score between images. Experimental results on various datasets demonstrate the effectiveness of our approach.

Person re-identification is an open and challenging problem in computer vision. Existing re-identification approaches focus on optimal methods for features matching (e.g., metric learning approaches) or study the inter-camera transformations of such features. These methods hardly ever pay attention to the problem of visual ambiguities shared between the first ranks. In this paper, we focus on such a problem and introduce an unsupervised ranking optimization approach based on discriminant context information analysis. The proposed approach refines a given initial ranking by removing the visual ambiguities common to first ranks. This is achieved by analyzing their content and context information. Extensive experiments on three publicly available benchmark datasets and different baseline methods have been conducted. Results demonstrate a remarkable improvement in the first positions of the ranking. Regardless of the selected dataset, state-of-the-art methods are strongly outperformed by our method.

Object, action, or scene representations that are corrupted by noise significantly impair the performance of visual recognition. Typically, partial occlusion, clutter, or excessive articulation affects only a subset of all feature dimensions and, most importantly, different dimensions are corrupted in different samples. Nevertheless, the common approach to this problem in feature selection and kernel methods is to down-weight or eliminate entire training samples or the same dimensions of all samples. Thus, valuable signal is lost, resulting in suboptimal classification. Our goal is, therefore, to adjust the contribution of individual feature dimensions when comparing any two samples and computing their similarity. Consequently, per-sample selection of informative dimensions is directly integrated into kernel computation. The interrelated problems of learning the parameters of a kernel classifier and determining the informative components of each sample are then addressed in a joint objective function. The approach can be integrated into the learning stage of any kernel-based visual recognition problem and it does not affect the computational performance in the retrieval phase. Experiments on diverse challenges of action recognition in videos and indoor scene classification show the general applicability of the approach and its ability to improve learning of visual representations.

We propose a method that extends a given depth image into regions in 3D that are not visible from the point of view of the camera. The algorithm detects repeated 3D structures in the visible scene and suggests a set of 3D extension hypotheses, which are then combined together through a global 3D MRF discrete optimization. The recovered global 3D surface is consistent with both the input depth map and the hypotheses. A key component of this work is a novel 3D template matcher that is used to detect repeated 3D structure in the scene and to suggest the hypotheses. A unique property of this matcher is that it can handle depth uncertainty. This is crucial because the matcher is required to ``peek around the corner'', as it operates at the boundaries of the visible 3D scene where depth information is missing. The proposed matcher is fast and is guaranteed to find an approximation to the globally optimal solution. We demonstrate on real-world data that our algorithm is capable of completing a full 3D scene from a single depth image and can synthesize a full depth map from a novel viewpoint of the scene. In addition, we report results on an extensive synthetic set of 3D shapes, which allows us to evaluate the method both qualitatively and quantitatively.

In this paper, we target on the problem of estimating the statistic of pedestrian travel time within a period from an entrance to a destination in a crowded scene. Such estimation is based on the global distributions of crowd densities and velocities instead of complete trajectories of pedestrians, which cannot be obtained in crowded scenes. The proposed method is motivated by our statistical investigation into the correlations between travel time and global properties of crowded scenes. Active regions are created for each source-destination pair to model the probable walking regions over the corresponding source-destination traffic flow. Two sets of scene features are specially designed for modeling moving and stationary persons inside the active regions and their influences on pedestrian travel time. The estimation of pedestrian travel time provides valuable information for both crowd scene understanding and pedestrian behavior analysis, but was not sufficiently studied in literature. The effectiveness of the proposed pedestrian travel time estimation model is demonstrated through several surveillance applications, including dynamic scene monitoring, localization of regions blocking traffics, and detection of abnormal pedestrian behaviors. Many more valuable applications based on our method are to be explored in the future.

In this work, we investigate the problem of automatically inferring the lattice structure of near-regular textures (NRT) in real-world images. Our technique leverages the PatchMatch algorithm for finding k-nearest-neighbor (kNN) correspondences in an image. We use these kNNs to recover an initial estimate of the 2D wallpaper basis vectors, and seed vertices of the texture lattice. We iteratively expand this lattice by solving an MRF optimization problem. We show that we can discretize the space of good solutions for the MRF using the kNNs, allowing us to efficiently and accurately optimize the MRF energy function using the Particle Belief Propagation algorithm. We demonstrate our technique on a benchmark NRT dataset containing a wide range of images with geometric and photometric variations, and show that our method clearly outperforms the state of the art in terms of both texel detection rate and texel localization score.

Patch based image modeling has achieved a great success in low level vision such as image denoising. In particular, the use of image nonlocal self-similarity (NSS) prior, which refers to the fact that a local patch often has many nonlocal similar patches to it across the image, has significantly enhanced the denoising performance. However, in most existing methods only the NSS of input degraded image is exploited, while how to utilize the NSS of clean natural images is still an open problem. In this paper, we propose a patch group (PG) based NSS prior learning scheme to learn explicit NSS models from natural images for high performance denoising. PGs are extracted from training images by putting nonlocal similar patches into groups, and a PG based Gaussian Mixture Model (PG-GMM) learning algorithm is developed to learn the NSS prior. We demonstrate that, owe to the learned PG-GMM, a simple weighted sparse coding model, which has a closed-form solution, can be used to perform image denoising effectively, resulting in high PSNR measure, fast speed, and particularly the best visual quality among all competing methods.

We address a new partial person re-identification (re-id) problem, where only a partial observation of a person is available for matching across different non-overlapping camera views. This differs significantly from the conventional person re-id setting where it is assumed that the full body of a person is detected and aligned. To solve this more challenging and realistic re-id problem without the implicit assumption of manual body-parts alignment, we propose a matching framework consisting of 1) a local patch-level matching model based on a novel sparse representation classification formulation with explicit patch ambiguity modelling, and 2) a global part-based matching model providing complementary spatial layout information. Our framework is evaluated on a new partial person re-id dataset as well as two existing datasets modified to include partial person images. The results show that the proposed method outperforms significantly existing re-id methods as well as other partial visual matching methods.

We propose a new family of discrete energy minimization problems, which we call parsimonious labeling. Our energy function consists of unary potentials and high-order clique potentials. While the unary potentials are arbitrary, the clique potentials are proportional to the diversity of the set of unique labels assigned to the clique. Intuitively, our energy function encourages the labeling to be parsimonious, that is, use as few labels as possible. This in turn allows us to capture useful cues for important computer vision applications such as stereo correspondence and image denoising. Furthermore, we propose an efficient graph-cuts based algorithm for the parsimonious labeling problem that provides strong theoretical guarantees on the quality of the solution. Our algorithm consists of three steps. First, we approximate a given diversity using a mixture of a novel hierarchical Pn Potts model. Second, we use a divide-and-conquer approach for each mixture component, where each subproblem is solved using an efficient alpha-expansion algorithm. This provides us with a small number of putative labelings, one for each mixture component. Third, we choose the best putative labeling in terms of the energy value. Using both synthetic and standard real datasets, we show that our algorithm significantly outperforms other graph-cuts based approaches.