CS 552/652 Speech Recognition with Hidden Markov Models Summer 2009 Oregon Health & Science University School of Science & Engineering Division of Biomedical Computer Science Center for Spoken Language Understanding John-Paul H

语音识别HMM算法中的解码功能的程序，其中是根据观测序列和已知模型去求状态序列-Speech recognition HMM algorithm decoding process, which is to seek observer status in accordance with the sequence and the known sequence of the model

Another common KWS method is Acoustic KWS. Using this method, the engine does not attempt to transcribe the entire stream of speech. Like the LVCSR-based method, this method employs the Viterbi search.