arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2012.13442eess.AScs.SD
收藏

多通道多帧ADL-MVDR目标语音分离

Multi-channel Multi-frame ADL-MVDR for Target Speech Separation

Indiana University(印第安纳大学);Tencent AI Lab(腾讯AI Lab)

Zhuohuang Zhang, Yong Xu, Meng Yu, Shi-Xiong Zhang, Lianwu Chen, Donald S. Williamson, Dong Yu

摘要:

已经提出了许多基于纯神经网络的语音分离方法来提高客观评估分数,但是它们经常引入对现代自动语音识别(ASR)系统有害的非线性失真。通常采用最小方差无失真响应(MVDR)滤波器来去除非线性失真,然而,传统的基于神经掩模的MVDR系统仍然导致相对高水平的残余噪声。此外,在与神经网络的联合训练期间,MVDR解决方案中涉及的矩阵求逆有时在数值上是不稳定的。在这项研究中,我们提出了一种用于目标语音分离的多通道多帧(MCMF)全深度学习(ADL)-MVDR方法,该方法扩展了我们初步的多通道ADL-MVDR方法。所提出的MCMF ADL-MVDR系统解决了线性和非线性失真。时空互相关也充分利用在所提出的方法。所提出的系统进行评估,使用普通话视听语料库,并与几个国家的最先进的方法进行比较。实验结果表明,我们提出的系统在不同的情况下,并在几个客观的评价指标,包括ASR性能的优越性。

英文摘要:

Many purely neural network based speech separation approaches have been proposed to improve objective assessment scores, but they often introduce nonlinear distortions that are harmful to modern automatic speech recognition (ASR) systems. Minimum variance distortionless response (MVDR) filters are often adopted to remove nonlinear distortions, however, conventional neural mask-based MVDR systems still result in relatively high levels of residual noise. Moreover, the matrix inverse involved in the MVDR solution is sometimes numerically unstable during joint training with neural networks. In this study, we propose a multi-channel multi-frame (MCMF) all deep learning (ADL)-MVDR approach for target speech separation, which extends our preliminary multi-channel ADL-MVDR approach. The proposed MCMF ADL-MVDR system addresses linear and nonlinear distortions. Spatio-temporal cross correlations are also fully utilized in the proposed approach. The proposed systems are evaluated using a Mandarin audio-visual corpus and are compared with several state-of-the-art approaches. Experimental results demonstrate the superiority of our proposed systems under different scenarios and across several objective evaluation metrics, including ASR performance.

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。

↑