FPGA based emotions recognition from speech signals

Abstract

Emotion recognition from speech signals has abundant applications in daily life. Particularly in speech-based human machine interaction it is used for improving the naturalness. Speech based emotion recognition is done in two steps namely Gender Recognition and Emotion Recognition. Gender Recognition will give the information about the gender (Male or Female) of the speaker and Emotion Recognition extracts the emotion (happy, sad, angry, and lazy etc.) of the speaker. In emotion recognition step back propagation algorithm under ANN is used as a classifier for classifying the emotions. In this paper, proposes emotion recognition from speech signal using Artificial Neural Networks (ANN) and implemented on FPGA device. The results using ANN are compared with the existing method and observed that ANN has lesser space utilization and improved speed than LDA.