Dynamic curve estimation for visual tracking.

摘要

In computer vision, visual tracking can be simply described as the action of locating/ detecting a given object in an image sequence by means of an algorithm. Tracking is often trivial for humans to perform, thanks to a highly effective pair of visual sensors, the use of other senses and superior cognitive capabilities. On the other hand, automated visual tracking is difficult for a variety of reasons; the primary of which include: significant clutter, non-trivial camera motion, occlusions, imaging noise and imperfect classification models. In this thesis, the visual tracking problem is tackled as a target contour estimation problem in the face of corrupted measurements.;The major aim of this thesis is to design robust recursive curve filters for the purpose of accurate visual contour-based tracking. The state-space representation adopted comprises of a group component and a shape component describing the rigid motion and the non-rigid shape deformation respectively; filtering strategies on each component are then decoupled. Due to the infinite dimension of the shape manifold, there is not a unique filtering update model for the shape component. Shapes being often described implicitly as the iso-contours of higher dimensional functions, the filtering strategy depends on the choice of the embedding function.;This thesis considers two implicit shape descriptors, a classification probability field and the traditional signed distance function, and aims to develop an optimal probabilistic contour observer and locally optimal curve filters. For the former, introducing a novel probabilistic shape description simplifies the filtering problem on the infinite-dimensional space of closed curves to a series of point-wise filtering tasks. The definition and justification of a novel update model suited to the shape space, the derivation of the filtering equations and the relation to Kalman filtering are studied. In addition to the temporal consistency provided by the filtering, extensions involving distributed filtering methods are considered in order to maintain spatial consistency. For the latter, locally optimal closed curve filtering strategies involving curve velocities are explored. The introduction of a local, linear description for planar curve variation and curve uncertainty enables the derivation of a mechanism for estimating the optimal gain associated to the curve filtering process, given quantitative uncertainty levels.;Experiments on synthetic and real sequences of images validate the filtering designs. While the techniques presented in this thesis are applied to planar curves, they can be extended to deal with the 3D cases involving surfaces.