Automatic face tracking and identity verification
Leila Essannouni, E. Ibn Elhaj, Driss Aboutajdine · 2007
The main objective of this work is to automate face tracking and verify the identity of the person. In fact the most pure algorithms of tracking has two major problems, namely the initialization problem and the lost track problem. To avoid these problems. The initialization is done by real time face detector. The verification of the presence of the searched face is done using a new technique of template matching based on robust correlation of the orientation of the second derivatives. The tracking is done using the sum of squared difference (SSD). The modules of face detection and SSD face tracker work orthogonally so that when one fails the other succeeds. The proposed framework is able to track and verify face identity of a person looking toward the camera.