工业场景下基于深度学习的语音去噪方法研究  

Research on Speech Denoising Methods Based on Deep Learning in Industrial Scenarios

作　　者：张辉 ZHANG Hui(Shandong Kunyun Enterprise Management Co.,Ltd.,Jinan 250011,China)

机构地区：[1]山东鲲云企业管理有限公司,山东济南250011

出　　处：《电声技术》2024年第8期48-50,共3页Audio Engineering

摘　　要：针对工业场景下的语音信号去噪问题,研究一种基于深度学习的语音去噪方法。先分析工业环境中的语音噪声特点,再研究基于长短期记忆(Long Short-Term Memory,LSTM)的语音去噪方法,并构建相应的损失函数,以提高去噪效果。实验结果表明,所提方法在重建误差方面明显优于标准的LSTM方法,有效提高了语音信号的品质和清晰度,可为工业场景下的语音通信系统的性能提升提供重要的理论基础和实用方法。This article focuses on the problem of speech signal denoising in industrial scenarios and studies a deep learning based method.Firstly,this article analyzes the characteristics of speech noise in industrial environments;Subsequently,the focus was on researching speech denoising methods based on Long Short-Term Memory(LSTM),and corresponding loss functions were constructed to improve the denoising effect.The experimental results show that the proposed method is significantly superior to the standard LSTM method in terms of reconstruction error,effectively improving the quality and clarity of speech signals.This study provides an important theoretical basis and practical method for improving the performance of speech communication systems in industrial scenarios.

关 键 词：深度学习 语音信号 去噪 损失函数 

分 类 号：TN912.3[电子电信—通信与信息系统] TP18[电子电信—信息与通信工程]