Facial Expression Recognition Algorithm Based on PCNN

A facial expression recognition method using the ignition position sequence (IPS) of pulse coupled neural network (PCNN) is proposed in this paper. First the method obtains normalized standard images by pre-processing the face images; and then uses PCNN model to extract the oscillation time sequences (OTS) of the face image, face identification is implemented by calculating the cosine distance of OTS and average OTS of all identities. Finally, face expression classification of testing face images is processed by using the IPS of PCNN. Experimental results in JAFFE expression database show that the method in this paper can efficiently distinguish the different expressions of the same identity, and achieve a higher recognition rate. Keywordsfacial expression recognition; pulse coupled neural network; oscillation time sequence