Knowledge Commons of Institute of Automation,CAS
Microphone Array Post-filter for Target Speech Enhancement Without a Prior Information of Point Interferers | |
Li GJ(李冠君); Liang S(梁山); Nie S(聂帅); Liu WJ(刘文举) | |
2020 | |
会议名称 | Interspeech |
会议日期 | 2020-10 |
会议地点 | shanghai |
摘要 | The post-filter for microphone array speech enhancement can effectively suppress noise including point interferers. However, the suppression of point interferers relies on the accurate estimation of the number and directions of point interferers, which is a difficult task in practical situations. In this paper, we propose a post-filtering algorithm, which is independent of the number and directions of point interferers. Specifically, we assume that the point interferers are continuously distributed at each direction of the plane but the probability of the interferer occurring at each direction is different in order to calculate the spatial covariance matrix of the point interferers. Moreover, we assume that the noise is additive and uncorrelated with the target signal to obtain the power spectral densities (PSDs) of the target signal and noise. Finally, the proposed post-filter is calculated using the estimated PSDs. Experimental results prove that the proposed post-filtering algorithm is superior to the comparative algorithms in the scenarios where the number and directions of point interferers are not accurately estimated |
七大方向——子方向分类 | 语音识别与合成 |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/44844 |
专题 | 多模态人工智能系统全国重点实验室_智能交互 |
作者单位 | Institute of Automation, Chinese Academy of Sciences |
第一作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Li GJ,Liang S,Nie S,et al. Microphone Array Post-filter for Target Speech Enhancement Without a Prior Information of Point Interferers[C],2020. |
条目包含的文件 | ||||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
20interspeech_后滤波.pd(639KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论