Temporal Discrete Cosine Transform for speech emotion recognition

Temporal Discrete Cosine Transform (TDCT) features have shown good performance in the speaker verification task, and in this paper we utilize them in speech emotion recognition. Tests were conducted on a Serbian emotional speech database, using Neural Networks (NN) as a classifier and Mel-Frequency Cepstral Coefficients (MFFC) as a reference feature set. Even though MFCC is one of the most employed techniques in emotion recognition, our results show that the TDCT features outperform MFCCs (with the first and second derivation) with any number of hidden nodes in the network, hence proving as an excellent starting feature set for recognizing emotions in South Slavic languages.