首页|苗语语音音节自适应切分算法

苗语语音音节自适应切分算法

扫码查看
原文链接
  • 万方数据
  • 维普
语音分割是苗语语音基础研究的难点和热点问题,其本质是苗语语音音节与沉默段(静音、噪音)之间边界模糊问题,目前相关研究成果较少.针对苗语语音音节分割边界模糊问题,以时域特征分割获得初始的语音边界,通过构建语音音节边界优化评价函数模型,将音节与沉默段之间边界模糊问题转化为音节真实边界和算法预测边界的误差极小化问题,以精英策略保留算法最佳优化分割边界种群个体,提取苗语语音音节与沉默段间的精准边界.实验结果表明,所提方法能显著提高苗语语音音节自适应边界搜索能力.此外,所提自适应分割算法在语音分割模糊边界分割方面显著优于典型的语音分割算法.
Adaptive Segmentation Algorithm for Hmong Language Syllable
Speech segmentation is a difficult and prominent issue in the basic research on Hmong speech.Its essence is the boundary ambiguity between Hmong speech syllables and silent segments(silence and noise).At present,there are few related re-search results.In order to solve the ambiguity problem of Hmong language syllable segmentation boundary,the time domain feature segmentation was performed on the obtained initial speech boundary.By constructing the optimization evaluation function model for speech syllable boundaries,the boundary ambiguity problem between syllables and silent segments was transformed into the error mini-mization between the real syllable boundary and the algorithmic predicted boundary.The elite strategy was used to preserve the best op-timized boundary population and extract the precise boundary between the syllable and the silent segment of Hmong speech.Experimen-tal results demonstrate that the proposed algorithm significantly improved the adaptive boundary search capability of Hmong speech syl-lables.Additionally,the experimental results show that the proposed adaptive segmentation algorithm out performs typical speech seg-mentation algorithms in handling the challenge of ambiguous boundaries in speech segmentation.

hmong speech segmentationtime domain featureshmong phonetic syllableelite strategyself-adaptation

冯夫健、吴磊、谭棉、蔡姗、张学文、王林

展开 >

贵州民族大学数据科学与信息工程学院,贵阳 550025

贵州省模式识别与智能系统重点实验室,贵阳 550025

贵阳人文科技学院大数据与信息工程学院,贵阳 550025

苗语语音分割 时域特征 苗语语音音节 精英策略 自适应

贵州省科技计划贵州省科技计划贵州省科技计划贵州省科技计划贵州省青年科技人才成长项目贵州省青年科技人才成长项目贵州省青年科技人才成长项目贵州省教育厅自然科学研究项目贵州省教育厅自然科学研究项目贵州省教育厅自然科学研究项目贵州省模式识别与智能系统重点实验室开放基金

黔科合基础-ZK[2022]一般195黔科合基础-ZK[2023]一般143黔科合平台人才-ZCKJ[2021]007黔科合基础-ZK[2024]一般493黔教合KY字[2021]104黔教合KY字[2021]113黔教合KY字[2021]110黔教技[2023]061号黔教技[2023]012号黔教技[2022]015号GZMUKL[2022]KF01

2024

10.12404/j.issn.1671-1815.2305356

科学技术与工程
中国技术经济学会

科学技术与工程

CSTPCD北大核心
影响因子：0.338
ISSN：1671-1815
年,卷(期)：2024.24(14)