MVNET: MEMORY ASSISTANCE AND VOCAL REINFORCEMENT NETWORK FOR SPEECH ENHANCEMENT

作     者：Wang, Jianrong Li, Xiaomin Li, Xuewei Yu, Mei Fang, Qiang Liu, Li 

作者机构：College of Intelligence and Computing Tianjin University Tianjin China Institute of Linguistics Chinese Academy of Social Sciences Beijing China Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Shenzhen China 

出 版 物：《arXiv》 (arXiv)

年 卷 期：2022年

核心收录：

主　　题：Speech enhancement 

摘      要：Speech enhancement improves speech quality and promotes the performance of various downstream tasks. However, most current speech enhancement work was mainly devoted to improving the performance of downstream automatic speech recognition (ASR), only a relatively small amount of work focused on the automatic speaker verification (ASV) task. In this work, we propose a MVNet consisted of a memory assistance module which improves the performance of downstream ASR and a vocal reinforcement module which boosts the performance of ASV. In addition, we design a new loss function to improve speaker vocal similarity. Experimental results on the Libri2mix dataset show that our method outperforms baseline methods in several metrics, including speech quality, intelligibility, and speaker vocal similarity et al. © 2022, CC0.