删除或更新信息，请邮件至freekaoyan#163.com(#换成@)

采用骨导语音自适应的语句分割方法*

本站小编 Free考研考试/2022-01-02
-->
苗晓孔,张雄伟.采用骨导语音自适应的语句分割方法*[J].,2019,38(1):68-75
采用骨导语音自适应的语句分割方法*
The adaptive speech segmentation method based on bone conduction voice
投稿时间：2018-04-12修订日期：2019-01-01
中文摘要:
为了满足含噪语音语句分割需求,也为了解决含噪语句或某些低信噪比环境下传统气导语句分割算法分割效果差,分割准确度低且算法自适应性弱等问题,提出一种基于骨导语音自适应的分段双门限语音分割方法。将骨导语音和气导语音同步采集,获取抗噪性能更好的骨导语音,然后在融合过零率与短时能量中引入随机动态阈值的自适应方法进行端点检测,最后利用分段双门限和语音聚类等手段实现语音分割,提高语音分割算法的鲁棒性。通过实验验证了所提算法的有效性和可行性,同时与其他语音分割算法进行了对比,证明本文所提分割算法精度更高,效果更好。
英文摘要:
In order to solve the problem of poor segmentation efficiency, low accuracy and weak adaptability of traditional gas guided speech segmentation algorithm under complex noise or low SNR, a segmented double threshold speech segmentation method based on bone guided speech is proposed. Firstly, the bone conduction and air conduction voice are collected synchronously to acquire the bone conduction speech with better anti-noise performance. Then the adaptive detection method based on random dynamic threshold is introduced in the fusion zero-crossing rate and short-time energy. Finally, the speech segmentation is realized by means of segmented double threshold and speech clustering, and the robustness of the speech segmentation algorithm is improved. The validity and feasibility of the proposed algorithm are verified by experiments. At the same time, compared with other speech segmentation algorithms, the proposed segmentation algorithm is proved to be more accurate and effective.
DOI：10.11684/j.issn.1000-310X.2019.01.010
中文关键词:骨导语音语音分割分段双门限语音聚类
英文关键词:Bone guided speechspeech segmentationsegmented double thresholdspeech clustering
基金项目:国家自然科学基金项目（61471394）
作者单位E-mail
苗晓孔陆军工程大学miao_xk@163.com
张雄伟陆军工程大学xwzhang9898@163.com
摘要点击次数:987
全文下载次数:657
查看全文查看/发表评论下载PDF阅读器
相关附件：修改说明118036原稿修改说明1修改说明2修改说明2修改说明1附件1作者变更附件1附件2
关闭








PDF全文下载地址:

http://yysx.cnjournals.cn/ch/reader/create_pdf.aspx?file_no=18036-1&flag=1&journal_id=yysx&year_id=2019
相关话题/陆军工程大学 中文 英文 实验 环境

Free考研考试FreeKaoYan.Com
欢迎来到Free考研考试，"为实现人生的Free而奋斗"

© 2020 FreeKaoYan! . All rights reserved.