Consonant/vowel segmentation for Mandarin syllable recognition

✍ Scribed by Ming-Tzaw Lin; Ching-Kuen Lee; Chin-Yi Lin

• Book ID: 102566829
• Publisher: Elsevier Science
• Year: 1999
• Tongue: English
• Weight: 308 KB
• Volume: 13
• Category: Article
• ISSN: 0885-2308
• DOI: 10.1006/csla.1999.0121

✦ Synopsis

A new system for the automatic segmentation and labeling of Mandarin speech is presented. The system is capable of labeling speech generated without requiring extensive linguistic knowledge or large training databases. In this work, a new approach using fuzzy implication was used to design a consonant/vowel segmentation method with a high accuracy rate and robustness to background noise for Mandarin syllable recognition systems. To find the abrupt spectral difference change of adjacent frames, we used the Lukasiewicz implication and the Gains implication to design a spectral distance measuring rule, which disregards small spectral differences and enhances large spectral differences. After measurement, the largest spectral difference of adjacent frames determines the segmentation of each consonant/vowel. The experimental results show that the accuracy of the obtained automatic segmentation is comparable to that of a human expert. The performance of our system appears to compare favorably with that of other systems.