arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2009.01941eess.AScs.SD
收藏

基于自注意的密集CNN时域语音增强

Dense CNN with Self-Attention for Time-Domain Speech Enhancement

The Ohio State University(俄亥俄州立大学)

Ashutosh Pandey, DeLiang Wang

摘要:

时域语音增强由于能够同时增强语音的幅度和相位,近年来变得越来越流行。在这项工作中,我们提出了一个密集的卷积网络(DCN)与自注意力的语音增强在时域。DCN是一种基于编码器和解码器的架构,具有跳过连接。编码器和解码器中的每一层包括密集块和注意模块。密集块和注意力模块有助于使用特征重用,增加网络深度和最大上下文聚合的组合进行特征提取。此外,我们揭示了以前未知的问题与增强语音的频谱幅度的基础上的损失。为了缓解这些问题,我们提出了一种新的损失的增强语音和预测噪声的幅度的基础上。即使所提出的损失仅基于幅度,由噪声预测施加的约束确保损失增强幅度和相位。实验结果表明,DCN训练与建议的损失大大优于其他国家的最先进的因果和非因果语音增强的方法。

英文摘要:

Speech enhancement in the time domain is becoming increasingly popular in recent years, due to its capability to jointly enhance both the magnitude and the phase of speech. In this work, we propose a dense convolutional network (DCN) with self-attention for speech enhancement in the time domain. DCN is an encoder and decoder based architecture with skip connections. Each layer in the encoder and the decoder comprises a dense block and an attention module. Dense blocks and attention modules help in feature extraction using a combination of feature reuse, increased network depth, and maximum context aggregation. Furthermore, we reveal previously unknown problems with a loss based on the spectral magnitude of enhanced speech. To alleviate these problems, we propose a novel loss based on magnitudes of enhanced speech and a predicted noise. Even though the proposed loss is based on magnitudes only, a constraint imposed by noise prediction ensures that the loss enhances both magnitude and phase. Experimental results demonstrate that DCN trained with the proposed loss substantially outperforms other state-of-the-art approaches to causal and non-causal speech enhancement.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑