Initialization and lost track recovery performance analysis of face features tracking

Abstract

Face feature tracking is an important element for detection and interpretation of human face expression and for further human mood-based interactions. This paper discusses the current status in face modeling techniques and performs a performance analysis. The analysis concerns the initialization and it also includes the lost track recovery for an Active Appearance Model-based face tracking algorithm.