A local time Fukunage-Koontz transformation for speech emotional recognition

Dimension reduction is widely used in the domains of speech emotional recognition. Considering the underlying time structure of speech signals, we propose a local time Fukunage-koontz transformation (LTFKT) to contain more discriminative information during emotional recognition. The goal of LTFKT is to maximize/ minimize the eigenvalues of covariance in different classes simultaneously. It can be seen that FKT is a special case of LTFKT from analysis. The results show that the method in this paper can improve recognition rate effectively.