Time and memory efficient viterbi decoding for LVCSR using a precompiled search network.

Daniel Willett, Erik McDermott, Yasuhiro Minami, Shigeru Katagiri

Browse the full Interspeech paper archive.