Speech recognition using temporally connected kernels in mixture density hidden Markov models.

Panu Somervuo

Browse the full ICASSP paper archive.