首页|基于自注意力机制的卷积循环网络语音降噪

基于自注意力机制的卷积循环网络语音降噪

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
由于对非平稳噪声进行估计是提高含噪语音降噪效果的重要影响因素,因此利用卷积模块提高单帧含噪语音所包含的信息,并通过Transformer中的自注意力机制模块,使模型能够更加精确区分含噪语音中的噪声部分和语音部分,从而使转置卷积模块更加高效的完成语音降噪.针对Noisex-92 噪声库中的 15 种噪声,分别应用LSTM网络、卷积循环网络和基于通道注意力机制的卷积循环网络模型进行对比分析,同时对测试集含噪语音进行降噪处理.实验结果表明,经过所提出的基于自注意力机制的卷积循环网络降噪后的语音在PESQ和STOI评分上均有较大提高,语谱图显示有效减少了噪声的残留.
Convolutional Recurrent Network Speech Denoising Based on Self-Attention Mechanism
Since the estimation of non-stationary noise is an important factor to improve the noise reduction effect of noisy speech,we used the convolution module to improve the information contained in single frame of noisy speech,and through the self-attention mechanism module in Transformer,enabled the model to distinguish the noise part from the speech part more accurately,so that the transpose convolution module more efficiently completed the speech noise reduction.LSTM network,convolutional loop network and convolutional loop network based on channel attention mechanism were used to compare and analyze 15 kinds of noise in Noisex-92 library.Meanwhile,noise reduction was performed for the noisy speech in the test set.The experimental results show that the proposed convolutional loop net-work based on self-attention mechanism has a great improvement in both PESQ and STOI scores,and the spectrogram display effectively reduces the residual noise.

Speech noise reductionNon-stationary noiseSelf-attention mechanismDeep learning

徐浩森、姜囡、齐志坤

展开 >

中国刑事警察学院公安信息技术与情报学院,辽宁沈阳 110854

证据科学教育部重点实验室(中国政法大学),北京 100088

语音降噪 非平稳噪声 自注意力机制 深度学习

证据科学教育部重点实验室(中国政法大学)开放基金辽宁省科技厅联合开放基金机器人学国家重点实验室开放基金中国刑事警察学院重大计划培育项目教育部重点项目辽宁省自然科学基金公安学科基础理论研究创新计划中央高校基本科研业务费专项公安学科基础理论研究创新计划

2021KFKT092020-KF-12-113242019010E-AQGABQ202027102019-ZD-016832420190102022XKGJ0110

2024

计算机仿真
中国航天科工集团公司第十七研究所

计算机仿真

CSTPCD
影响因子：0.518
ISSN：1006-9348
年,卷(期)：2024.41(4)
  • 21