HMM-based speech recognition using state-dependent, linear transforms on Mel-warped DFT features.

Chengalv Rathinavelu, Li Deng

Browse the full ICASSP paper archive.