Human Emotion Recognition from Speech in Audio Physical Features

Abstract

Emotion is the part of human life which highly influences decision making compared to other things. The human emotion is not only conveyed by the body language but also through speech. Speech delivers the maximum information regarding the feelings of human. Nowadays, still there are many scenarios that these emotions are not being recognized properly. This is why this area has taken the attention of the researcher to explore. This fact has motivated us to propose a methodology to discriminate human emotions precisely. The proposed methodology has selected a set of features which carry more information about emotional feelings. The audio files database has been preprocessed. Physical features means both time-domain and frequency-domain features have been extracted from the preprocessed data. The files have been classified based on the feature. Some well-known classifier has been adopted for classification to achieve better outcome.