DEVELOPMENT OF AN ALGORITHM FOR EXTRACTION OF SPECTRAL, PROSODIC AND MFCC FEATURES

Аннотация

This paper presents an in-depth analysis of the main groups of features derived from the speech signal – spectral, prosodic, and Mel-Frequency Cepstral Coefficients (MFCC) — and proposes algorithms for their extraction. One of the key stages in speech-based emotion recognition is the accurate extraction of emotional features from the signal. Regardless of the type of classification model employed, its accuracy and robustness directly depend on the quality of the used features. Therefore, this study focuses primarily on the effective extraction of such features