Emotion Recognition from Acted Assamese Speech

Abstract

Emotions play important roles in expressing feelings as it tend to make people acts differently. Determining emotions of the speaker is less complicated if we are facing him/her rather than from voice independently such as conversation in telephone. However it would be a great achievement if we able to detect with what emotion a speaker is speaking just by listening to the voice. This project is a small step towards it and we basically are focusing on determining emotions through the recorded speech and developing the prototype system. The ability to detect human emotion from their speech is going to be a great addition in the field of human-robot interaction. The aim of the work is to build an emotion recognition system using Mel-frequency cepstral coefficients (MFCC) and Gaussian mixture model (GMM). In this work four emotional states happy, sad, angry and neutral are taken for classification of emotions. Here we have considered only 10 speakers, 7 male and 3 female, all belonging to upper Assam region and all speak in the same accent. The experiments are performed for only speaker dependent and text independent case.