Emotion Recognition in Speech using Back Propagation Algorithm

Abstract

Speech emotion detection refers to discovering the speech category based on the training and testing to the database provided. This research work has been classified in four sections namely SAD, HAPPY, FEAR and AGGRESSIVE. There are two major sections in this research work namely Training and Testing. The training has been done on the basis of wave files provided for every group. Features have been extracted for all groups and have been saved into the database. The testing section classifies the training set of data with the help of BACK PROPAGATION NEURAL NETWORK (BPN) classifier and SEQUENTIAL MINIMAL OPTIMIZATION (SMO) classifier. The results of the BACK PROPAGATION NEURAL NETWORK CLASSIFIER have been found superior in terms of classification accuracy.