ICCV 2015 论文列表

The objective of this work is human pose estimation in videos, where multiple frames are available. We investigate a ConvNet architecture that is able to benefit from temporal context by combining information across the multiple frames using optical flow. To this end we propose a network architecture with the following novelties: (i) a deeper network than previously investigated for regressing heatmaps; (ii) spatial fusion layers that learn an implicit spatial model; (iii) optical flow is used to align heatmap predictions from neighbouring frames; and (iv) a final parametric pooling layer which learns to combine the aligned heatmaps into a pooled confidence map. We show that this architecture outperforms a number of others, including one that uses optical flow solely at the input layers, one that regresses joint coordinates directly, and one that predicts heatmaps without spatial fusion. The new architecture outperforms the state of the art by a large margin on three video pose estimation datasets, including the very challenging Poses in the Wild dataset, and outperforms other deep methods that don't use a graphical model on the single-image FLIC benchmark (and also Chen & Yuille and Tompson et al. in the high precision region).

Convolutional neural networks (CNNs) have recently been very successful in a variety of computer vision tasks, especially on those linked to recognition. Optical flow estimation has not been among the tasks CNNs succeeded at. In this paper we construct CNNs which are capable of solving the optical flow estimation problem as a supervised learning task. We propose and compare two architectures: a generic architecture and another one including a layer that correlates feature vectors at different image locations. Since existing ground truth data sets are not sufficiently large to train a CNN, we generate a large synthetic Flying Chairs dataset. We show that networks trained on this unrealistic data still generalize very well to existing datasets such as Sintel and KITTI, achieving competitive accuracy at frame rates of 5 to 10 fps.

The Flickr30k dataset has become a standard benchmark for sentence-based image description. This paper presents Flickr30k Entities, which augments the 158k captions from Flickr30k with 244k coreference chains linking mentions of the same entities in images, as well as 276k manually annotated bounding boxes corresponding to each entity. Such annotation is essential for continued progress in automatic image description and grounded language understanding. We present experiments demonstrating the usefulness of our annotations for text-to-image reference resolution, or the task of localizing textual entity mentions in an image, and for bidirectional image-sentence retrieval. These experiments confirm that we can further improve the accuracy of state-of-the-art retrieval methods by training with explicit region-to-phrase correspondence, but at the same time, they show that accurately inferring this correspondence given an image and a caption remains really challenging.

Detecting fine-grained subtle changes among a scene is critically important in practice. Previous change detection methods, focusing on detecting large-scale significant changes, cannot do this well. This paper proposes a feasible end-to-end approach to this challenging problem. We start from active camera relocation that quickly relocates camera to nearly the same pose and position of the last time observation. To guarantee detection sensitivity and accuracy of minute changes, in an observation, we capture a group of images under multiple illuminations, which need only to be roughly aligned to the last time lighting conditions. Given two times observations, we formulate fine-grained change detection as a joint optimization problem of three related factors, i.e., normal-aware lighting difference, camera geometry correction flow, and real scene change mask. We solve the three factors in a coarse-to-fine manner and achieve reliable change decision by rank minimization. We build three real-world datasets to benchmark fine-grained change detection of misaligned scenes under varied multiple lighting conditions. Extensive experiments show the superior performance of our approach over state-of-the-art change detection methods and its ability to distinguish real scene changes from false ones caused by lighting variations.

The visual perception of object affordances has emerged as a useful ingredient for building powerful computer vision and robotic applications. In this paper we introduce a novel approach to reason about liquid containability - the affordance of containing liquid. Our approach analyzes container objects based on two simple physical processes: the Fill and Transfer of liquid. First, it reasons about whether a given 3D object is a liquid container and its best filling direction. Second, it proposes directions to transfer its contained liquid to the outside while avoiding spillage. We compare our simplified model with a common fluid dynamics simulation and demonstrate that our algorithm makes human-like choices about the best directions to fill containers and transfer liquid from them. We apply our approach to reason about the containability of several real-world objects acquired using a consumer-grade depth camera.

L_0 gradient minimization can be applied to an input signal to control the number of non-zero gradients. This is useful in reducing small gradients generally associated with signal noise, while preserving important signal features. In computer vision, L_0 gradient minimization has found applications in image denoising, 3D mesh denoising, and image enhancement. Minimizing the L_0 norm, however, is an NP-hard problem because of its non-convex property. As a result, existing methods rely on approximation strategies to perform the minimization. In this paper, we present a new method to perform L_0 gradient minimization that is fast and effective. Our method uses a descent approach based on region fusion that converges faster than other methods while providing a better approximation of the optimal L_0 norm. In addition, our method can be applied to both 2D images and 3D mesh topologies. The effectiveness of our approach is demonstrated on a number of examples.

In this paper, we consider the problem of estimating the gaze direction of a person from a low-resolution image. Under this condition, reliably extracting facial features is very difficult. We propose a novel head pose estimation algorithm based on compressive sensing. Head image patches are mapped to a large feature space using the proposed extensive, yet efficient filter bank. The filter bank is designed to generate sparse responses of color and gradient information, which can be compressed using random projection, and classified by a random forest. Extensive experiments on challenging datasets show that the proposed algorithm performs favorably against the state-of-the-art methods on head pose estimation in low-resolution images degraded by noise, occlusion, and blurring.

We propose a family of structured matrices to speed up orthogonal projections for high-dimensional data commonly seen in computer vision applications. In this, a structured matrix is formed by the Kronecker product of a series of smaller orthogonal matrices. This achieves O(dlogd) computational complexity and O(logd) space complexity for d-dimensional data, a drastic improvement over the standard unstructured projections whose computational and space complexities are both O(d^2). The proposed structured matrices are applicable to a number of application domains, and are faster and more compact than other structured matrices used in the past. We also introduce an efficient learning procedure for optimizing such matrices in a data dependent fashion. We demonstrate the significant advantages of the proposed approach in solving the approximate nearest neighbor (ANN) image search problem with both binary embedding and quantization. We find that the orthogonality plays a very important role in solving ANN problem, since the random orthogonal Kronecker projection has already provided promising performance. Comprehensive experiments show that the proposed approach can achieve similar or better accuracy as the existing state-of-the-art but with significantly less time and memory.

We present a method to continuously blend between multiple facial performances of an actor, which can contain different facial expressions or emotional states. As an example, given sad and angry video takes of a scene, our method empowers the movie director to specify arbitrary weighted combinations and smooth transitions between the two takes in post-production. Our contributions include (1) a robust nonlinear audio-visual synchronization technique that exploits complementary properties of audio and visual cues to automatically determine robust, dense spatiotemporal correspondences between takes, and (2) a seamless facial blending approach that provides the director full control to interpolate timing, facial expression, and local appearance, in order to generate novel performances after filming. In contrast to most previous works, our approach operates entirely in image space, avoiding the need of 3D facial reconstruction. We demonstrate that our method can synthesize visually believable performances with applications in emotion transition, performance correction, and timing control.

In this paper, we propose a method for the robust and efficient computation of multi-frame optical flow in an expressive sequence of facial images. We formulate a novel energy minimisation problem for establishing dense correspondences between a neutral template and every frame of a sequence. We exploit the highly correlated nature of human expressions by representing dense facial motion using a deformation basis. Furthermore, we exploit the even higher correlation between deformations in a given input sequence by imposing a low-rank prior on the coefficients of the deformation basis, yielding temporally consistent optical flow. Our proposed model-based formulation, in conjunction with the inverse compositional strategy and low-rank matrix optimisation that we adopt, leads to a highly efficient algorithm for calculating facial flow. As experimental evaluation, we show quantitative experiments on a challenging novel benchmark of face sequences, with dense ground truth optical flow provided by motion capture data. We also provide qualitative results on a real sequence displaying fast motion and occlusions. Extensive quantitative and qualitative comparisons demonstrate that the proposed method outperforms state-of-the-art optical flow and dense non-rigid registration techniques, whilst running an order of magnitude faster.

We propose a novel easy-to-implement stroke detector based on an efficient pixel intensity comparison to surrounding pixels. Stroke-specific keypoints are efficiently detected and text fragments are subsequently extracted by local thresholding guided by keypoint properties. Classification based on effectively calculated features then eliminates non-text regions. The stroke-specific keypoints produce 2 times less region segmentations and still detect 25% more characters than the commonly exploited MSER detector and the process is 4 times faster. After a novel efficient classification step, the number of regions is reduced to 7 times less than the standard method and is still almost 3 times faster. All stages of the proposed pipeline are scale- and rotation-invariant and support a wide variety of scripts (Latin, Hebrew, Chinese, etc.) and fonts. When the proposed detector is plugged into a scene text localization and recognition pipeline, a state-of-the-art text localization accuracy is maintained whilst the processing time is significantly reduced.

The baselines of a document page are a set of virtual horizontal and parallel lines, to which the printed contents of document, e.g., text lines, tables or inserted photos, are aligned. Accurate baseline extraction is of great importance in the geometric correction of curved document images. In this paper, we propose an efficient method for accurate extraction of these virtual visual cues from a curved document image. Our method comes from two basic observations that the baselines of documents do not intersect with each other and that within a narrow strip, the baselines can be well approximated by linear segments. Based upon these observations, we propose a curvilinear projection based method and model the estimation of curved baselines as a constrained sequential optimization problem. A dynamic programming algorithm is then developed to efficiently solve the problem. The proposed method can extract the complete baselines through each pixel of document images in a high accuracy. It is also scripts insensitive and highly robust to image noises, non-textual objects, image resolutions and image quality degradation like blurring and non-uniform illumination. Extensive experiments on a number of captured document images demonstrate the effectiveness of the proposed method.

Natural image modeling plays a key role in many vision problems such as image denoising. Image priors are widely used to regularize the denoising process, which is an illposed inverse problem. One category of denoising methods exploit the priors (e.g., TV, sparsity) learned from external clean images to reconstruct the given noisy image, while another category of methods exploit the internal prior (e.g., self-similarity) to reconstruct the latent image. Though the internal prior based methods have achieved impressive denoising results, the improvement of visual quality will become very difficult with the increase of noise level. In this paper, we propose to exploit image external patch prior and internal self-similarity prior jointly, and develop an external patch prior guided internal clustering algorithm for image denoising. It is known that natural image patches form multiple subspaces. By utilizing Gaussian mixture models (GMMs) learning, image similar patches can be clustered and the subspaces can be learned. The learned GMMs from clean images are then used to guide the clustering of noisypatches of the input noisy images, followed by a low-rank approximation process to estimate the latent subspace for image recovery. Numerical experiments show that the proposed method outperforms many state-of-the-art denoising algorithms such as BM3D and WNNM.

Catadioptric imaging systems use curved mirrors to capture wide fields of view. However, due to the curvature of the mirror, these systems tend to have very limited depth of field (DOF), with the point spread function (PSF) varying dramatically over the field of view and as a function of scene depth. In recent years, focal sweep has been used extensively to extend the DOF of conventional imaging systems. It has been shown that focal sweep produces an integrated point spread function (IPSF) that is nearly space-invariant and depth-invariant, enabling the recovery of an extended depth of field (EDOF) image by deconvolving the captured focal sweep image with a single IPSF. In this paper, we use focal sweep to extend the DOF of a catadioptric imaging system. We show that while the IPSF is spatially varying when a curved mirror is used, it remains quasi depth-invariant over the wide field of view of the imaging system. We have developed a focal sweep system where mirrors of different shapes can be used to capture wide field of view EDOF images. In particular, we show experimental results using spherical and paraboloidal mirrors.

Causal relationships can often be found in visual object tracking between the motions of the camera and that of the tracked object. This object motion may be an effect of the camera motion, e.g. an unsteady handheld camera. But it may also be the cause, e.g. the cameraman framing the ob- ject. In this paper we explore these relationships, and pro- vide statistical tools to detect and quantify them; these are based on transfer entropy and stem from information the- ory. The relationships are then exploited to make predic- tions about the object location. The approach is shown to be an excellent measure for describing such relationships. On the VOT2013 dataset the prediction accuracy is increased by 62 % over the best non-causal predictor. We show that the location predictions are robust to camera shake and sud- den motion, which is invaluable for any tracking algorithm and demonstrate this by applying causal prediction to two state-of-the-art trackers. Both of them benefit, Struck gain- ing a 7 % accuracy and 22 % robustness increase on the VTB1.1 benchmark, becoming the new state-of-the-art.

Despite recent progress, reconstructing outdoor scenes in 3D from movable platforms remains a highly difficult endeavour. Challenges include low frame rates, occlusions, large distortions and difficult lighting conditions. In this paper, we leverage the fact that the larger the reconstructed area, the more likely objects of similar type and shape will occur in the scene. This is particularly true for outdoor scenes where buildings and vehicles often suffer from missing texture or reflections, but share similarity in 3D shape. We take advantage of this shape similarity by localizing objects using detectors and jointly reconstructing them while learning a volumetric model of their shape. This allows us to reduce noise while completing missing surfaces as objects of similar shape benefit from all observations for the respective category. We evaluate our approach with respect to LIDAR ground truth on a novel challenging suburban dataset and show its advantages over the state-of-the-art.

We present a novel approach to 3D reconstruction which is inspired by the human visual system. This system unifies standard appearance matching and triangulation techniques with higher level reasoning and scene understanding, in order to resolve ambiguities between different interpretations of the scene. The types of reasoning integrated in the approach includes recognising common configurations of surface normals and semantic edges (e.g. convex, concave and occlusion boundaries). We also recognise the coplanar, collinear and symmetric structures which are especially common in man made environments.

We present an approach to modeling ear-to-ear, high-quality texture from one or more partial views of a face with possibly poor resolution and noise. Our approach is example-based in that we reconstruct texture with patches from a database composed of previously seen faces. A 3D morphable model is used to establish shape correspondence between the observed data across views and training faces. The database is built on the mesh surface by segmenting it into uniform overlapping patches. Texture patches are selected by belief propagation so as to be consistent with neighbors and with observations in an appropriate image formation model. We also develop a variant that is insensitive to light and camera parameters, and incorporate soft symmetry constraints. We obtain textures of higher quality for degraded views as small as 10 pixels wide, than a standard model fitted to non-degraded data. We further show applications to super-resolution where we substantially improve quality compared to a state-of-the-art algorithm, and to texture completion where we fill in missing regions and remove facial clutter in a photorealistic manner.

Recently several generalizations of the popular latent structural SVM framework have been proposed in the literature. Broadly speaking, the generalizations can be divided into two categories: (i) those that predict the output variables while either marginalizing the latent variables or estimating their most likely values; and (ii) those that predict the output variables by minimizing an entropy-based uncertainty measure over the latent space. In order to aid their application in computer vision, we study these generalizations with the aim of identifying their strengths and weaknesses. To this end, we propose a novel prediction criterion that includes as special cases all previous prediction criteria that have been used in the literature. Specifically, our framework's prediction criterion minimizes the Aczel and Daroczy entropy of the output. This in turn allows us to design a learning objective that provides a unified framework (UF) for latent structured prediction. We develop a single optimization algorithm and empirically show that it is as effective as the more complex approaches that have been previously employed for latent structured prediction. Using this algorithm, we provide empirical evidence that lends support to prediction via the minimization of the latent space uncertainty.