Application of Vector Quantization in Emotion Recognition from Human Speech

Recognition of emotions from speech is a complex task that is furthermore complicated by the fact that there is no unambiguous answer to what the "correct" emotion is for a given speech sample. In this paper, we discuss emotion classification of a well known German database consisting of 6 basic emotions: sadness, boredom, neutral, fear, happiness, and anger using Mel frequency Cepstral Coefficients (MFCCs). A concern with MFCC is the large number of features. We discuss the use of LBG-VQ algorithm to minimize the amount of data to be handled. At last, emotion classification is done using Euclidean distance, Manhattan distance and Chebyshev distance of the codebooks between neutral state and other emotional states for the same sample.