A fast and accurate algorithm for facial feature segmentation

A facial feature segmentation algorithm Sor head and shoulder sequences is proposed. The method is based on simple spatial and temporal heuristics techniques for the purpose of practical implementation. Facial feature are segmented using two common cues i) regular and high image intensity variation in temporal domain due to various motion like talking and moving, ii) high edge density around the feature regions. Simple motion detection method and a generic spatial operator are applied to achieve the required segmentation. The paper addresses issue of how to combine both information to obtain final results and also assess the segmentation performance.