Human emotion identification using feed forward neural network with backpropagation and bayesian regularized backpropagation algorithm

Abstract

In this work, initially the human face will be detected. Then the facial features will be extracted and classified into different expressions. Here two types of algorithm viz. used in Feed Forward neural network(FFNN), ie., Backpropagation(BP) Algorithm and Bayesian Regularization Algorithm. After evaluating Bayesian regularized Backpropagation Algorithm (BR) is found to be better suited for automatic facial expression recognition than Backpropagation algorithm(BP), and the performance is evaluate using various metrics.