Facial Expression Classification System with Emotional Back Propagation Neural Network

Abstract

Facial expression recognition is also gaining interest among the researchers because of its inevitable advantages in image retrieval which can be extended many fields like medicine, artificial intelligence, robotics and neural networks. So it is one of the hot topic for researchers. Existing methods such as PCA, LDA, LPP etc. with Euclidian distance classifier are popular. Neural network classifiers are also used along with above for classification. In this paper, the pattern averaging and PCA are used for feature extraction. In this work, feed forward neural network with added emotional coefficients (EBPNN) for facial expression classification is being proposed. The network is trained with back propagation algorithm. The results are compared with normal feed forward neural network with back propagation. The proposed algorithm is producing better results over the existing.