Marker-less based animation generation using motion control

Abstract

A highly active research area in Computer Science is that of automatic vision-based human motion capture. Human motion capture is invaluable in many areas including human computer interfaces, surveillance possibly to monitor unusual behaviour, and last but not least computer animation [l]. The aim of this dissertation is to automatically generate a 2D cartoon model from a skeleton model by using a video camera. This means that marker-less tracking and object detection where also two other aims. It is important to note that the system proposed here is a monocular one i.e. it uses only one camera.