Speech Emotion Recognition Classifiers: A Literature Review
- Automatic speech emotion recognition plays a significant role in human–computer interaction. The basic goals of the automatic emotion recognition are to understand and retrieve emotions from individual’s speech utterance. The main key issues of speech emotion recognition systems are to prepare an appropriate dataset, selection of suitable feature sets, and design of a proper classification method. The selected feature vectors are stored in the database and later fed to classifier to detect the emotions by comparing the vectors from the trained data and test data vectors. In this paper we review the various classification techniques based on support vector machines, Gaussian Mixture Models, Hidden Markov Models, Artificial Neural Network, k-Nearest Neighbors and Naïve Bayes. We also try to present some literary survey work to show the use of classifiers by different authors.
Keywords - Speech emotion classifier; SVM; GMM; HMM; ANN; KNN; Naïve Bayes