Abstract: In the field of speech-based emotion recognition, Mel-Frequency Cepstral Coefficients (MFCCs) are widely used as representative acoustic features. Recent approaches have transformed MFCCs ...