Title: Fuzzy Expert System for Continuous Speech Recognition
Author: Hajin Yu, Yung Hwan Oh, Yoichi Yamashita and Riichiro Mizoguchi
Reference: Proc. of '93 Korea/Japan Joint Conference on Expert Systems, Seoul, pp.951-965.
In designing a speech recognition expert system, there are some aspects to be considered - segmentation method, recognition unit, structure of the rule, and the user aids. To avoid the difficulties in segmentation, an irregular unit based on spectral transition measure is proposed. Frame is used as the structure of the speech recognition rules. The structure provides the user an easy way of making rules, and it enables the building of an automatic rule generator. Fuzzy linguistic variables are used in representating rules. The rule-generating cycle which includes rule generator, rule tester with the error reporter, and state descriptor with other modules, could save much time in making rules and could get a high performance. The experimental result and comparison with the SPREX system - the basis of this system - are described.