Facial expression recognition using local binary patterns with different distance measures

Abstract

Facial expression recognition is a well-known activity in the domain of human–computer interaction and computer vision. In this work, we have applied face detection algorithm on the images to get the facial part only; then, we have used local binary pattern (LBP) operator to get the facial features. Finally, to match the test image with the different expressions, various distance measures which are Euclidian distance, taxicab distance, chessboard distance, Bray–Curtis distance and chi-square distance have been applied. The maximum facial expression recognition rate of Bray–Curtis distance measure reaches 92.85% for person-dependent expression recognition, which is better than other distance measures. The experiments were performed on JAFFE which is a standard dataset, and result shows that the facial expression recognition with LBP and Bray–Curtis for person-dependent recognition is an effective method. © Springer India 2014.