[1]王珊,谷城..基于归一化自相关语音筛选的鲁棒性说话人确认研究[J].郑州大学学报(工学版),2009,30(03):141-144.[doi:10.3969/j.issn.1671-6833.2009.03.036]
Wang Shan,Valley City.A study of robust speaker confirmation based on normalized autocorrelated speech screening[J].Journal of Zhengzhou University (Engineering Science),2009,30(03):141-144.[doi:10.3969/j.issn.1671-6833.2009.03.036]
点击复制
基于归一化自相关语音筛选的鲁棒性说话人确认研究(
)
《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]
- 卷:
-
30卷
- 期数:
-
2009年03期
- 页码:
-
141-144
- 栏目:
-
- 出版日期:
-
1900-01-01
文章信息/Info
- Title:
-
A study of robust speaker confirmation based on normalized autocorrelated speech screening
- 作者:
-
王珊; 谷城.
-
河南大学,物理与电子学院,河南,开封,475000, 河南大学,物理与电子学院,河南,开封,475000
- Author(s):
-
Wang Shan; Valley City
-
-
- 关键词:
-
归一化自相关; 语音筛选; 鲁棒性; 说话人确认
- Keywords:
-
- DOI:
-
10.3969/j.issn.1671-6833.2009.03.036
- 文献标志码:
-
A
- 摘要:
-
语音在受到加性噪声污染时,不同部分受噪声污染的程度存在差异,保留受污染严重的语音将会对说话人确认系统产生负面影响,笔者提出了一种基于归一化自相关的鲁棒性语音筛选方法.通过归一化自相关舍弃受污染严重的语音.实验表明,通过归一化自相关舍掉受污染严重的语音,能提高噪声环境下的说话人确认性能.
- Abstract:
-
When speech is polluted by additive noise, the degree of noise pollution of different parts is different, and retaining heavily polluted speech will have a negative impact on the speaker confirmation system. Discard heavily contaminated speech by normalized autocorrelation. Experiments show that normalized autocorrelation can improve the confirmation performance of speakers in noisy environment by discarding heavily polluted speech.
更新日期/Last Update:
1900-01-01