Visual tracking using dynamic transition in groups of affine transformed templates

Abstract

This paper presents a view-based visual tracking system. View-based systems are more general than the other approaches which require analysis of 2D image features and matching with 3D models. For the image correlation, we use a visual tracking hardware based on a block matching algorithm. One problem of the block matching is how to cope with changes of the template's appearance in the 3D environment. Consequently, the system generates affine transformed templates corresponding to various orientations of the target image plane. We use geodesic domes not only for uniform distribution of the 3D orientations of the template image plane, but also for dynamic transition in groups of the candidate templates for the next search step. Experimental results demonstrate the usefulness of the proposed system.