Investigating deep neural network based transforms of robust audio features for LVCSR.

Enrico Bocchieri, Dimitrios Dimitriadis

Browse the full ICASSP paper archive.