Abstract
This paper proposes a method of real-time detection and tracking of facial feature points from the facial image sequence taken by a web camera, and the automatic real-time synthesis of facial caricature animation. Firstly the pupils are detected in the initial frame by using the circular separation filter. Next the area of each facial part is localized from the positions of pupils, and facial feature points are detected in each area by image processing methods such as YIQ transform, Canny edge operators, and so on. These detection processes are carried out in every frame and facial caricature animation is generated based on the detected feature points. In order to carry out the process from capturing moving facial images to generation of facial caricature animation in real-time, GPGPU and parallel processing methods are introduced to accelerate the detection of facial feature points.
