SPEED IMPROVEMENT OF THE TIME-ASYNCHRONOUS ACOUSTIC FAST MATCH

Abstract

This paper describes an algorithm for improvement of the speed of a time-asynchronous fast match, which is a part of a stack-search based recognition system. This fast match uses a phonetic tree to represent the entire vocabulary of the recognizer. Evaluation of the tree (in a depth-first manner), can be done much more efficiently using the fact that under certain conditions, the results of branch evaluations can be used to approximate the scores of other branches of the tree.