A Direct-Concatenation Approach to Train Hidden Markov Models to Recognize the Highly Confusing Mandarin Syllables with Very Limited Training Data
Journal
IEEE Transactions on Speech and Audio Processing
Journal Volume
1
Journal Issue
1
Pages
113-119
Date Issued
1993
Author(s)
Abstract
syllables is very difficult because this vocabulary consists of 38 confusing sets, each of which can have as many as 19 syllables. The recognition of these 408 syllables becomes even more difficult when only very limited training data is available. In this paper, a special direct-concatenation approach to train hidden Markov models (HMM’s) to recognize these syllables with very limited training data is developed, in which each syllable is divided into INITIAL and FINAL parts and 408 right-context-dependent INITIAL HMM’s and 38 left-context-independent FINAL HMM’s are separately trained with the transition region carefully taken of, and then these INITIAL and FINAL HMM’s are directly concatenated to form syllable recognition. Experimental results show that this approach can utilize the very limited training data most efficiently and provide significant improvements in recognition performance. Although the results are obtained for Mandarin syllables, the approach is believed to be equally helpful for the recognition of other confusing vocabularies. The recognition of a total of 408 very confusing Mandarin. © 1993 IEEE
SDGs
Type
journal article