Simultaneously Tracking and Recognition of Facial Features and Facial Expressions

Abstract

The tracking and recognition of facial activities from images have attracted a great attention in computer vision field. Face activities are divided in the three different levels. First, the bottom level, facial feature points around every facial sections, i.e., eyebrow, mouth, etc., capture the face information point by point. Second, in the middle level, facial activity units, defined in the facial activity coding framework, illustrate the construction of a specific set of facial muscles, i.e., lid tightener, eyebrow raiser, and so forth. At last, in the top level, six types of facial expressions are illustrate the facial muscle movement and are usually used to display the human feeling states. Rather than the standard approaches, which in general just concentrate on a two levels of face activities, and tracking them independently, this paper presents a Advanced machine learning techniques are used to learn the model in view of both training information and subjective prior knowledge. Given the model and the estimates the facial movements, and each of the three individual levels of face activities are recognized at the same time through an expression engine. Experiments are performed to illustrate effectiveness of the above proposed model.