A study on different approaches of machine learning on Emotion Recognition from human speech

Abstract

This paper present the basic study of different approaches of machine learning towards emotion recognition from human speech. It presents the various approaches used for feature extraction, a set of different features, challenges towards speech recognition. The classifiers used for pattern matching with all its advantages and disadvantages. It is a survey on the existing approaches for recognizing emotions from human speech and accuracy rates achieved using different models and datasets by different authors.  It also focuses on the different applications of emotion recognition in different fields like diagnosing psychological disorders, improved human machine interaction and other real life scenarios etc.