CVPR 2015 Papers

We propose a robust algorithm to generate video segment proposals. The proposals generated by our method can start from any frame in the video and are robust to complete occlusions. Our method does not assume specific motion models and even has a limited capability to generalize across videos. We build on our previous least squares tracking framework, where image segment proposals are generated and tracked using learned appearance models. The innovation in our new method lies in the use of two efficient moves, the merge move and free addition, to efficiently start segments from any frame and track them through complete occlusions, without much additional computation. Segment size interpolation is used for effectively detecting occlusions. We propose a new metric for evaluating video segment proposals on the challenging VSB-100 benchmark and present state-of-the-art results. Preliminary results are also shown for the potential use of our framework to track segments across different videos.

In the field of saliency detection, many graph-based algorithms heavily depend on the accuracy of the pre-processed superpixel segmentation, which leads to significant sacrifice of detail information from the input image. In this paper, we propose a novel bottom-up saliency detection approach that takes advantage of both region-based features and image details. To provide more accurate saliency estimations, we first optimize the image boundary selection by the proposed erroneous boundary removal. By taking the image details and region-based estimations into account, we then propose the regularized random walks ranking to formulate pixel-wised saliency maps from the superpixel-based background and foreground saliency estimations. Experiment results on two public datasets indicate the significantly improved accuracy and robustness of the proposed algorithm in comparison with 12 state-of-the-art saliency detection approaches.

In this paper, we present a computationally efficient and non-parametric method for robust regression on manifolds. We apply our algorithm to the problem of correcting mislabeled examples from image collections with ordered (e.g., real-valued, ordinal) labels. Compared to related methods for robust regression, our method achieves superior denoising accuracy on a variety of data sets, with label corruption levels as high as 80%. For a diverse set of widely-used, large-scale, publicly-available data sets, our approach results in image labels that more accurately describe the associated images.

We present an approach to indoor scene reconstruction from RGB-D video. The key idea is to combine geometric registration of scene fragments with robust global optimization based on line processes. Geometric registration is error-prone due to sensor noise, which leads to aliasing of geometric detail and inability to disambiguate different surfaces in the scene. The presented optimization approach disables erroneous geometric alignments even when they significantly outnumber correct ones. Experimental results demonstrate that the presented approach substantially increases the accuracy of reconstructed scene models.

The estimation of multiple homographies between two piecewise planar views of a rigid scene is often assumed to be a solved problem. We show that contrary to popular opinion various crucial aspects of the task have not been adequately emphasised. We are motivated by a growing body of literature in robust multi-structure estimation that purports to solve the multi-homography estimation problem but in fact does not. We demonstrate that the estimation of multiple homographies is an ill-solved problem by deriving new constraints that a set of mutually compatible homographies must satisfy, and by showing that homographies estimated with prevailing methods fail to satisfy the requisite constraints on real-world data. We also explain why incompatible homographies imply inconsistent epipolar geometries. The arguments and experiments presented in this paper signal the need for a new generation of robust multi-structure estimation methods that have the capacity to enforce constraints on projective entities such as homography matrices.

This paper proposes a robust multi-image based blind face hallucination framework to super-resolve LR faces. The proposed framework first estimates both blurring kernel and transformations of multiple LR faces by robust deblurring and registration in PCA subspace. A patch-wise mixture of probabilistic PCA prior is then incorporated for face super-resolution. Previous work on face SR using PCA prior can be viewed as special cases of the framework. Experimental results in both simulated and real LR sequences demonstrate very promising performance of the proposed method.

This paper proposes a new framework for estimating the Manhattan Frame (MF) of an indoor scene from a single RGB-D image. Our technique formulates this problem as the estimation of a rotation matrix that best aligns the normals of the captured scene to a canonical world axes. By introducing sparsity constraints, our method can simultaneously estimate the scene MF, the surfaces in the scene that are best aligned to one of three coordinate axes, and the outlier surfaces that do not align with any of the axes. To test our approach, we contribute a new set of annotations to determine ground truth MFs in each image of the popular NYUv2 dataset. We use this new benchmark to experimentally demonstrate that our method is more accurate, faster, more reliable and more robust than the methods used in the literature. We further motivate our technique by showing how it can be used to address the RGB-D SLAM problem in indoor scenes by incorporating it into and improving the performance of a popular RGB-D SLAM method.

This paper deals with the trajectory estimation of a monocular calibrated camera evolving in a large unknown environment, also known as monocular visual simultaneous localization and mapping. The contribution of this paper is threefold: 1) We develop a new formalism that builds upon the so called Known Rotation Problem to robustly estimate submaps (parts of the camera trajectory and the unknown environment). 2) In order to obtain a globally consistent map (up to a scale factor), we propose a novel loopy belief propagation algorithm that is able to efficiently align a large number of submaps. Our approach builds a graph of relative 3D similarities (computed between the submaps) and estimates the global 3D similarities by passing messages through a super graph until convergence. 3) To render the whole framework more robust, we also propose a simple and efficient outlier removal algorithm that detects outliers in the graph of relative 3D similarities. We extensively demonstrate, on the TUM and KITTI benchmarks as well as on several other challenging video sequences, that the proposed method outperforms the state of the art algorithms.

Regularizing images under a guidance signal has been used in various tasks in computer vision and computational photography, particularly for noise reduction and joint upsampling. The aim is to transfer fine structures of guidance signals to input images, restoring noisy or altered structures. One of main drawbacks in such a data-dependent framework is that it does not handle differences in structure between guidance and input images. We address this problem by jointly leveraging structural information of guidance and input images. Image filtering is formulated as a nonconvex optimization problem, which is solved by the majorization-minimization algorithm. The proposed algorithm converges quickly while guaranteeing a local mini- mum. It effectively controls image structures at different scales and can handle a variety of types of data from different sensors. We demonstrate the flexibility and effectiveness of our model in several applications including depth super- resolution, scale-space filtering, texture removal, flash/non- flash denoising, and RGB/NIR denoising.

This paper addresses two issues hindering the advances in accurate image alignment. First, the performance of descriptor-based approaches to image alignment relies on the chosen descriptor, but the optimal descriptor typically varies from image to image, or even pixel to pixel. Second, the neighborhood structure for smoothness enforcement is usually predefined before alignment. However, object boundaries are often better discovered during alignment. The proposed approach tackles the two issues by adaptive descriptor selection and dynamic neighborhood construction. Specifically, we associate each pixel to be aligned with an affine transformation, and integrate the learning of the pixel-specific transformations into image alignment. The transformations serve as the common domain for descriptor fusion, since the local consensus of each descriptor can be estimated by accessing the corresponding affine transformation. It allows us to pick the most plausible descriptor for aligning each pixel. On the other hand, more object-aware neighborhoods can be produced by referencing the consistency between the learned affine transformations of neighboring pixels. The promising results on popular image alignment benchmarks manifests the effectiveness of our approach.

3D structure recovery from a collection of 2D images requires the estimation of the camera locations and orientations, i.e. the camera motion. For large, irregular collections of images, existing methods for the location estimation part, which can be formulated as the inverse problem of estimating n locations from noisy measurements of a subset of the pairwise directions, are sensitive to outliers in direction measurements. In this paper, we firstly provide a complete characterization of well-posed instances of the location estimation problem, by presenting its relation to the existing theory of parallel rigidity. For robust estimation of camera locations, we introduce a two-step approach, comprised of a pairwise direction estimation method robust to outliers in point correspondences between image pairs, and a convex program to maintain robustness to outlier directions. In the presence of partially corrupted measurements, we empirically demonstrate that our convex formulation can even recover the locations exactly. Lastly, we demonstrate the utility of our formulations through experiments on Internet photo collections.

While sparse coding on non-flat Riemannian manifolds has recently become increasingly popular, existing solutions either are dedicated to specific manifolds, or rely on optimization problems that are difficult to solve, especially when it comes to dictionary learning. In this paper, we propose to make use of kernels to perform coding and dictionary learning on Riemannian manifolds. To this end, we introduce a general Riemannian coding framework with its kernel-based counterpart. This lets us (i) generalize beyond the special case of sparse coding; (ii) introduce efficient solutions to two coding schemes; (iii) learn the kernel parameters; (iv) perform unsupervised and supervised dictionary learning in a much simpler manner than previous Riemannian coding methods. We demonstrate the effectiveness of our approach on three different types of non-flat manifolds, and illustrate its generality by applying it to Euclidean spaces, which also are Riemannian manifolds.

Compressive sensing(CS) has been exploited for hypespectral image(HSI) compression in recent years. Though it can greatly reduce the costs of computation and storage, the reconstruction of HSI from a few linear measurements is challenging. The underlying sparsity of HSI is crucial to improve the reconstruction accuracy. However, the sparsity of HSI is unknown in reality and varied with different noise, which makes the sparsity estimation difficult. To address this problem, a novel reweighted Laplace prior based hyperspectral compressive sensing method is proposed in this study. First, the reweighted Laplace prior is proposed to model the distribution of sparsity in HSI. Second, the latent variable Bayes model is employed to learn the optimal configuration of the reweighted Laplace prior from the measurements. The model unifies signal recovery, prior learning and noise estimation into a variational framework to infer the parameters automatically. The learned sparsity prior can represent the underlying structure of the sparse signal very well and is adaptive to the unknown noise, which improves the reconstruction accuracy of HSI. The experimental results on three hyperspectral datasets demonstrate the proposed method outperforms serveral state-of-the-art hyperspectral CS methods on the reconstruction accuracy.

We present a simple but powerful reinterpretation of kernelized locality-sensitive hashing (KLSH), a general and popular method developed in the vision community for performing approximate nearest-neighbor searches in an arbitrary reproducing kernel Hilbert space (RKHS). Our new perspective is based on viewing the steps of the KLSH algorithm in an appropriately projected space, and has several key theoretical and practical benefits. First, it eliminates the problematic conceptual difficulties that are present in the existing motivation of KLSH. Second, it yields the first formal retrieval performance bounds for KLSH. Third, our analysis reveals two techniques for boosting the empirical performance of KLSH. We evaluate these extensions on several large-scale benchmark image retrieval data sets, and show that our analysis leads to improved recall performance of at least 12%, and sometimes much higher, over the standard KLSH method.

In this paper, we present and experiment a novel approach for representing texture of 3D mesh manifolds using local binary patterns (LBP). Using a recently proposed framework [37], we compute LBP directly on the mesh surface, either using geometric or photometric appearance. Compared to its depth-image counterpart, our approach is distinguished by the following features: a) inherits the intrinsic advantages of mesh surface (e.g., preservation of the full geometry); b) does not require normalization; c) can accommodate partial matching. In addition, it allows early-level fusion of the geometry and photometric texture modalities. Through experiments conducted on two application scenarios, namely, 3D texture retrieval and 3D face recognition, we assess the effectiveness of the proposed solution with respect to state of the art approaches.

The goal of this paper is to enable a 3D "virtual-tour" of an apartment given a small set of monocular images of different rooms, as well as a 2D floor plan. We frame the problem as inference in a Markov Random Field which reasons about the layout of each room and its relative pose (3D rotation and translation) within the full apartment. This gives us accurate camera pose in the apartment for each image. What sets us apart from past work in layout estimation is the use of floor plans as a source of prior knowledge, as well as localization of each image within a bigger space (apartment). In particular, we exploit the floor plan to impose aspect ratio constraints across the layouts of different rooms, as well as to extract semantic information, e.g., the location of windows which are marked in floor plans. We show that this information can significantly help in resolving the challenging room-apartment alignment problem. We also derive an efficient exact inference algorithm which takes only a few ms per apartment. This is due to the fact that we exploit integral geometry as well as our new bounds on the aspect ratio of rooms which allow us to carve the space, significantly reducing the number of physically possible configurations. We demonstrate the effectiveness of our approach on a new dataset which contains over 200 apartments.

Most modern trackers typically employ a bounding box given in the first frame to track visual objects, where their tracking results are often sensitive to the initialization. In this paper, we propose a new tracking method, Reliable Patch Trackers (RPT), which attempts to identify and exploit the reliable patches that can be tracked effectively through the whole tracking process. Specifically, we present a tracking reliability metric to measure how reliably a patch can be tracked, where a probability model is proposed to estimate the distribution of reliable patches under a sequential Monte Carlo framework. As the reliable patches distributed over the image, we exploit the motion trajectories to distinguish them from the background. Therefore, the visual object can be defined as the clustering of homo-trajectory patches, where a Hough voting-like scheme is employed to estimate the target state. Encouraging experimental results on a large set of sequences showed that the proposed approach is very effective and in comparison to the state-of-the-art trackers. The full source code of our implementation will be publicly available.

Part-based models are one of the leading paradigms in visual recognition. In the absence of costly part annotations, associating and aligning different training instances of a part classifier and finding characteristic negatives is challenging and computationally demanding. To avoid this costly mining of training samples, we estimate separate generative models for negatives and positives and integrate them into a max-margin exemplar-based model. The generative model and a sparsity constraint on the correlation between spatially neighboring feature dimensions regularize the part filters during learning and improve their generalization to similar instances. To suppress inappropriate positive part samples, we project the classifier back into the image domain and penalize against deviations from the original exemplar image patch. The part filter is then optimized to i) discriminate against clutter, to ii) generalize to similar instances of the part, and iii) to yield a good reconstruction of the original image patch. Moreover, we propose an approximation for estimating the geometric margin so that learning large numbers of parts becomes feasible. Experiments show improved part localization, object recognition, and part-based reconstruction performance compared to popular exemplar-based approaches on PASCAL VOC

We study the problem of temporally consistent video post-processing. Previous post-processing algorithms usually either fail to keep high fidelity or fail to keep temporal consistency of output videos. In this paper, we observe experimentally that many image/video enhancement algorithms enforce a spatially consistent prior on the enhancement. More precisely, within a local region, the enhancement is consistent, i.e., pixels with the same RGB values will get the same enhancement values. Using this prior, we segment each frame into several regions and temporally-spatially adjust the enhancement of regions of different frames, taking into account fidelity, temporal consistency and spatial consistency. User study, objective measurement and visual quality comparisons are conducted. The experimental results demonstrate that our output videos can keep high fidelity and temporal consistency at the same time.