Speech emotion recognition using a fuzzy approach

Article type: Research Article

Authors: Ton-That, An H.^a | Cao, Nhan T.^{b; *}

Affiliations: [a] Faculty of Information Technology, University of Finance and Marketing, Ho Chi Minh City, Vietnam | [b] Faculty of Information Systems, University of Information Technology, Vietnam National University Ho Chi Minh City, Vietnam

Correspondence: [*] Corresponding author. N.T. Cao. E-mail: [email protected].

Abstract:  This paper introduces a fuzzy approach for classifying speech emotions in which a fuzzy inference system based on fuzzy associative memory (FAM-FIS) is used for recognizing speech emotions. Experiments on two databases of emotion speech Emo-DB in German and SAVEE in English, and feature of Mel-Frequency Cepstral Coefficients (MFCC) showed that the accuracy rates of the fuzzy inference system are better than that of Bayes and Support Vector Machine (SVM) on same kind of features and databases. Namely, with MFCC feature and 19 dimensions, Emo-DB is 74.31% and SAVEE is 97.29%.

Keywords: Speech emotion recognition, fuzzy inference systems, fuzzy associative memory, pattern recognition, fuzzy approach

DOI: 10.3233/JIFS-18594

Journal: Journal of Intelligent & Fuzzy Systems, vol. 36, no. 2, pp. 1587-1597, 2019