Facial Expression Recognition Using AAMICPF

Recently, many interests have been focused on the facial expression recognition research because of its importance in many applications area. In the computer vision area, the object recognition and the state recognition are very important and critical. Variety of researches have been done and proposed but those are very difficult to solve. We propose, in this paper, to use Active Appearance Model (AAM) with Particle filter for facial expression recognition system. AAM is very sensitive about initial shape. So we improve accuracy using Particle filter which is defined by the initial state to particles. Our system recognizes the facial expression using each criteria expression vector. We find better result than using basic AAM and 10% improvement has been made with AAA-IC.