AUTOMATIC SEGMENTATION AND LABELING BASED ON BOUNDARY TYPE FOR MANDARIN CHINESE SPEECH

摘要：

This paper presents an automatic segmentation and labeling method for mandarin Chinese speech synthesis corpus. In order to improve the accuracy of segmentation, two types of HMM models are utilized to produce the INITIAL/FINAL and syllable boundaries. Three feature detection algorithms are applied to boundary refinement for speech boundaries of voiced/unvoiced/silence. Experimental results show the proposed method can improve the performance of the segmentation system significantly.

关键词： Speech segmentation HMM Phoneme boundary-type V/U/S detection Measurement of the distance between segments

作者: PING-MU HUANG CHENG-LI SUN JUN GUO

作者单位: School of Information Engineering, Beijing University of Posts and Telecommunications, China

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 2497-2502

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）