Model tracking for video-based virtual reality

Abstract

This paper presents a technique for tracking complex objects (both polyhedral and smooth boundaries) in a monocular sequence. Our aim is to use this model tracking method in an augmented reality context to compute the pose of a real object to be able to register it with a synthetic one. A scalar score function for an object pose is defined, based on the local image gradient along the projected model boundaries. A local search is then carried out in the configuration space of the pose to maximize the score. This technique is robust to occlusions, since the whole object contour is used, not just a few control points. The proposed method is effective yet simple. No image feature extraction is necessary and no complex temporal evolution is used. Experimental results with a real sequence show the good performance of our technique.