First International Conference on Spoken Language Processing (ICSLP 90)

Kobe, Japan
November 18-22, 1990

Discrimination of Words in a Large Vocabulary Speech Recognition System

S. Datta (1), M. Al-Zabibi (2)

(1) Dept. of Electronic & Electrical Engineering, Loughborough University of Technology, Leics, UK
(2) Scientific Studies & Research Centre, Damascus, Syria

Word discrimination according to broad phonetic classes is presented in this paper. Different Phonetic classification strategies are used to describe large vocabulary lexicons. The phonetic description in these strategies varies from 2 to 17 phonetic classes. The statistical results show that about 83% of the 10,000 test Arabic words can be uniquely represented by using 7 broad phonetic classes for consonants and six classes for vowels. In this case, the maximum number of words having the same phonetic labelling is 6. The paper summarises the results of ten different phonetic classification schemes and discusses their implication for a large vocabulary speech recognition system. Distributions of vowels and consonant classes are also presented.

Full Paper

Bibliographic reference.  Datta, S. / Al-Zabibi, M. (1990): "Discrimination of words in a large vocabulary speech recognition system", In ICSLP-1990, 249-252.