Detection , tracking and registration of shapes with a vision system Student

Sensor fusion can be used to recover 3D pose and to analyze spatial configuration of object parts. The goal of this project is to develop a sensor fusion method using vision sensors (stereo cameras, color vision). This project addresses the visual tracking problem for a rigid-articulated object using a known geometrical and kinematical model. This model provides us with useful information about its motion constraints. For fast computation, the model is built with elliptical primitives. This project describes the implementation of a vision-based object tracker, which uses prior knowledge from an object of interest including color, shape and spatial configuration. The tracker uses multiple modalities derived from stereovision and color images. The object is first defined through an interactive interface that allows the user to build an ellipse-based model of it. Detection, tracking and registration of shapes with a vision system 6