Fast and Low Power Viterbi Search Engine Using Inverse Hidden Markov Model

Bo-Sung KIM  Jun-Dong CHO  

Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E87-A   No.3   pp.695-697
Publication Date: 2004/03/01
Online ISSN: 
DOI: 
Print ISSN: 0916-8508
Type of Manuscript: Special Section LETTER (Special Section on Applications and Implementations of Digital Signal Processing)
Category: Communication Theory and Systems
Keyword: 
VLSI,  HMM,  Viterbi search,  low-power,  

Full Text: PDF>>
Buy this Article




Summary: 
Viterbi search engine in speech recognition consumes many computation time and hardware resource for finding maximum likelihood in HMM (Hidden Markov Model). We propose a fast Viterbi search engine using IHMM (Inverse Hidden Markov Model). A benefit of this method is that we can remove redundant computation of path matrix. The power consumption and the computational time are reduced by 68.6% at the 72.9% increase in terms of the number of gates.