Recognition of Facial Expressions With Respect to Navarasas in Bharathanatyam Styles Using Neural Network.

Abstract

This research aims at developing an intelligent system which can recognize facial expressions related to human beings. Six universally accepted prominent emotions such as anger, fear, happiness, disgust, sadness and surprise along with neutral ones are recognized. In addition to this, emotions in Bharathanatyam, a classical dance style of south India is also studied in this paper by using neural network. Background subtraction method, face masking method and forward back propagation methods are used to train the neural network to recognize the faces. Results are found to be very encouraging. We hope this would open up a new window to the face recognition field.