Springer Science and Business Media LLC

The advancement of visual tracking has continuously been brought by deep learning models. Typically, supervised learning is employed to train these models with expensive labeled data. In order to reduce the workload of manual annotation and learn to track arbitrary objects, we propose an unsupervised learning method for visual tracking. The motivation of our unsupervised learning is that a robust tracker should be effective in bidirectional tracking. Specifically, the tracker is able to forward localize a target object in successive frames and backtrace to its initial position in the first frame. Based on such a motivation, in the training process, we measure the consistency between forward and backward trajectories to learn a robust tracker from scratch merely using unlabeled videos. We build our framework on a Siamese correlation filter network, and propose a multi-frame validation scheme and a cost-sensitive loss to facilitate unsupervised learning. Without bells and whistles, the proposed unsupervised tracker achieves the baseline accuracy of classic fully supervised trackers while achieving a real-time speed. Furthermore, our unsupervised framework exhibits a potential in leveraging more unlabeled or weakly labeled data to further improve the tracking accuracy.

In this article we provide a very robust algorithm to compute the position of the center of a sphere with known radius from one image by a calibrated camera. To our knowledge it is the first time that an exact sphere localization formula is published that only uses the (pixel) area and the ellipse center of the sphere image. Other authors either derived an approximation formula or followed the less robust and more time consuming procedure of fitting an ellipse through the detected edge pixels. Our method is analytic and deterministic, making use of the unique positive real tool of a cubic equation. We observe that the proposed area method is significantly more accurate and precise than an ellipse fitting method. Furthermore, we investigate in what conditions for sphere images the proposed exact method is preferable to the robust approximation method. These observations are validated by virtual, synthetic and real experiments.

This paper describes a new algorithm for disparity estimation using trinocular stereo. The three cameras are placed in a right angled configuration. A graph is then constructed whose nodes represent the individual pixels and whose edges are along the epipolar lines. Using the well known uniqueness and ordering constraint for pair by pair matches simultaneously, a path with the least matching cost is found using dynamic programming and the disparity filled along the path. This process is repeated iteratively until the disparity at all the pixels are filled up. To demonstrate the effectiveness of our approach, we present results from real world images and compare it with the traditional line by line stereo using dynamic programming.

In this paper we analyze in some detail the geometry of a pair of cameras, i.e., a stereo rig. Contrarily to what has been done in the past and is still done currently, for example in stereo or motion analysis, we do not assume that the intrinsic parameters of the cameras are known (coordinates of the principal points, pixels aspect ratio and focal lengths). This is important for two reasons. First, it is more realistic in applications where these parameters may vary according to the task (active vision). Second, the general case considered here, captures all the relevant information that is necessary for establishing correspondences between two pairs of images. This information is fundamentally projective and is hidden in a confusing manner in the commonly used formalism of the Essential matrix introduced by Longuet-Higgins (1981). This paper clarifies the projective nature of the correspondence problem in stereo and shows that the epipolar geometry can be summarized in one 3×3 matrix of rank 2 which we propose to call the Fundamental matrix. After this theoretical analysis, we embark on the task of estimating the Fundamental matrix from point correspondences, a task which is of practical importance. We analyze theoretically, and compare experimentally using synthetic and real data, several methods of estimation. The problem of the stability of the estimation is studied from two complementary viewpoints. First we show that there is an interesting relationship between the Fundamental matrix and three-dimensional planes which induce homographies between the images and create unstabilities in the estimation procedures. Second, we point to a deep relation between the unstability of the estimation procedure and the presence in the scene of so-called critical surfaces which have been studied in the context of motion analysis. Finally we conclude by stressing the fact that we believe that the Fundamental matrix will play a crucial role in future applications of three-dimensional Computer Vision by greatly increasing its versatility, robustness and hence applicability to real difficult problems.

We present a visual recognition system for fine-grained visual categorization. The system is composed of a human and a machine working together and combines the complementary strengths of computer vision algorithms and (non-expert) human users. The human users provide two heterogeneous forms of information object part clicks and answers to multiple choice questions. The machine intelligently selects the most informative question to pose to the user in order to identify the object class as quickly as possible. By leveraging computer vision and analyzing the user responses, the overall amount of human effort required, measured in seconds, is minimized. Our formalism shows how to incorporate many different types of computer vision algorithms into a human-in-the-loop framework, including standard multiclass methods, part-based methods, and localized multiclass and attribute methods. We explore our ideas by building a field guide for bird identification. The experimental results demonstrate the strength of combining ignorant humans with poor-sighted machines the hybrid system achieves quick and accurate bird identification on a dataset containing 200 bird species.