Combination of temporal domain SVD based speech enhancement and GMM based speech estimation for ASR in noise - evaluation on the AURORA2 task -.

Masakiyo Fujimoto, Yasuo Ariki

Browse the full Interspeech paper archive.