In many retrieval, object recognition and wide baseline stereo methods, correspondences of interest points (distinguished regions) are commonly established by matching compact descriptors such as SIFTs. We show that a subsequent cosegmentation process coupled with a quasi-optimal sequential decision process leads to a correspondence verification procedure that (i) has high precision (is highly discriminative) (ii) has good recall and (iii) is fast. The sequential decision on the correctness of a correspondence is based on simple statistics of a modified dense stereo matching algorithm. The statistics are projected on a prominent discriminative direction by SVM. Wald’s sequential probability ratio test is performed on the SVM projection computed on progressively larger cosegmented regions. We show experimentally that the proposed Sequential Correspondence Verification (SCV) algorithm significantly outperforms the standard correspondence selection method based on SIFT distance ratios on challenging matching problems.

This paper presents a novel local image descriptor that is robust to general image deformations. A limitation with traditional image descriptors is that they use a single support region for each interest point. For general image deformations, the amount of deformation for each location varies and is unpredictable such that it is difficult to choose the best scale of the support region. To overcome this difficulty, we propose to use multiple support regions of different sizes surrounding an interest point. A feature vector is computed for each support region, and the concatenation of these feature vectors forms the descriptor for this interest point. Furthermore, we propose a new similarity measure model, Local-to-Global Similarity (LGS) model, for point matching that takes advantage of the multi-size support regions. Each support region acts as a ’weak ’ classifier and the weights of these classifiers are learned in an unsupervised manner. The proposed approach is evaluated on a number of images with real and synthetic deformations. The experiment results show that our method outperforms existing techniques under different deformations. 1.

We present a novel variational approach for simultaneous segmentation of two images of the same object taken from different viewpoints. Due to noise, clutter and occlusions, neither of the images contains sufficient information for correct object-background partitioning. The evolving object contour in each image provides a dynamic prior for the segmentation of the other object view. We call this process mutual segmentation. The foundation of the proposed method is a unified level-set framework for region and edge based segmentation, associated with a shape similarity term. The suggested shape term incorporates the semantic knowledge gained in the segmentation process of the image pair, accounting for excess or deficient parts in the estimated object shape. Transformations, including planar projectivities, between the object views are accommodated by a registration process held concurrently with the segmentation. The proposed segmentation algorithm is demonstrated on a variety of image pairs. The Homography between each of the image pairs is estimated and its accuracy is evaluated.

We propose an approach to identify and segment objects from scenes that a person (or robot) encounters in Activities of Daily Living (ADL). Images collected in those cluttered scenes contain multiple objects. Each image provides only a partial, possibly very different view of each object. An object instance discovery program must be able to link pieces of visual information from multiple images and extract the consistent patterns. Most papers on unsupervised discovery of object models are concerned with object categories. In contrast, this paper aims at identifying and extracting regions corresponding to specific object instances, e.g., two different laptops in the laptop category. By focusing on specific instances, we enforce explicit constraints on geometric consistency (such as scale, orientation), and appearance consistency (such as color, texture and shape). Using multiple segmentations as the basic building block, our program processes a noisy “soup ” of segments and extracts object models as groups of mutually consistent segments. Our approach was tested on three different types of image sets: two from indoor ADL environments and one from Flickr.com. The results demonstrate robustness of our program to severe clutter, occlusion, changes of viewpoint and interference from irrelevant images. Our approach achieves significant improvement over with two existing methods. 1.

Abstract. We exploit the recurrence of structures at different locations, orientations and scales in an image to perform denoising. While previous methods based on “nonlocal filtering ” identify corresponding patches only up to translations, we consider more general similarity transformations. Due to the additional computational burden, we break the problem down into two steps: First, we extract similarity invariant descriptors at each pixel location; second, we search for similar patches by matching descriptors. The descriptors used are inspired by scale-invariant feature transform (SIFT), whereas the similarity search is solved via the minimization of a cost function adapted from local denoising methods. Our method compares favorably with existing denoising algorithms as tested on several datasets. 1

We introduce a new approach to reconstructing accurate camera geometry and 3D models for urban structures in a holistic fashion, i.e., without relying on extraction or matching of traditional local features such as points and edges. Instead, we use semi-global or global features based on transform invariant low-rank textures, which are ubiquitous in urban scenes. Modern high-dimensional optimization techniques enable us to accurately and robustly recover precise and consistent camera calibration and scene geometry from single or multiple images of the scene. We demonstrate how to construct 3D models of large-scale buildings from sequences of multiple large-baseline uncalibrated images that conventional SFM systems do not apply. 1.

We present a tracking algorithm based on motion analysis of regional affine invariant image features. The tracked object is represented with a probabilistic occupancy map. Using this map as support, regional features are detected and probabilistically matched across frames. The motion of pixels is then established based on the feature motion. The object occupancy map is in turn updated according to the pixel motion consistency. We describe experiments to measure the sensitivities of our approach to inaccuracy in initialization, and compare it with other approaches. 1.

Abstract not found

( * denotes joint first author.)