arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.06994eess.AS
收藏

ADL-MVDR:用于目标语音分离的全深度学习MVDR波束形成器

ADL-MVDR: All deep learning MVDR beamformer for target speech separation

Zhuohuang Zhang, Yong Xu, Meng Yu, Shi-Xiong Zhang, Lianwu Chen, Dong Yu

摘要:

语音分离算法通常用于将目标语音从其他干扰源中分离出来。然而,基于纯神经网络的语音分离系统经常导致非线性失真,这对于自动语音识别(ASR)系统是有害的。传统的基于掩模的最小方差无失真响应(MVDR)波束形成器可以用来最小化失真,但伴随着高水平的残留噪声。此外,矩阵运算(例如,矩阵求逆)在与神经网络联合训练时有时在数值上不稳定。在本文中,我们提出了一种新的全深度学习MVDR框架,其中矩阵求逆和特征值分解由两个递归神经网络(RNN)代替,以同时解决这两个问题。该方法利用RNN预测的帧间波束形成权值,在保持目标语音不失真的同时,大大降低了残余噪声。该系统是评估的普通话视听语料库和比较几个国家的最先进的(SOTA)语音分离系统。实验结果表明,该方法在几个客观指标和ASR准确性的优越性。

英文摘要:

Speech separation algorithms are often used to separate the target speech from other interfering sources. However, purely neural network based speech separation systems often cause nonlinear distortion that is harmful for automatic speech recognition (ASR) systems. The conventional mask-based minimum variance distortionless response (MVDR) beamformer can be used to minimize the distortion, but comes with high level of residual noise. Furthermore, the matrix operations (e.g., matrix inversion) involved in the conventional MVDR solution are sometimes numerically unstable when jointly trained with neural networks. In this paper, we propose a novel all deep learning MVDR framework, where the matrix inversion and eigenvalue decomposition are replaced by two recurrent neural networks (RNNs), to resolve both issues at the same time. The proposed method can greatly reduce the residual noise while keeping the target speech undistorted by leveraging on the RNN-predicted frame-wise beamforming weights. The system is evaluated on a Mandarin audio-visual corpus and compared against several state-of-the-art (SOTA) speech separation systems. Experimental results demonstrate the superiority of the proposed method across several objective metrics and ASR accuracy.

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。

↑