A speech enhancement method combining beamforming with RNN for hearing aids

IF 0.5 Q4 ENGINEERING, MULTIDISCIPLINARY

Journal of Computational Methods in Sciences and Engineering Pub Date : 2023-05-31 DOI:10.3233/jcm-226897

Zhiqian Qiu, Fei Chen, Junyu Ji

{"title":"A speech enhancement method combining beamforming with RNN for hearing aids","authors":"Zhiqian Qiu, Fei Chen, Junyu Ji","doi":"10.3233/jcm-226897","DOIUrl":null,"url":null,"abstract":"Speech enhancement is essential for hearing aids. In recent years, many speech enhancement methods based on deep learning have been proven to be effective. However, these speech enhancement methods rarely consider limited hardware resources and have difficulty meeting real-time requirements, which is very important for hearing aids. To solve the above problems, we propose a method that combines beamforming and speech enhancement methods based on deep learning. Beamforming is used to filter background noise and reduce the complexity of noise. Additionally, a new filter bank used in hearing aids is adopted to reduce the complexity of the system. The system was deployed and tested in resource-constrained hearing aids. The effectiveness of the method was verified by objective experiments using standard evaluation indicators. The results showed that the power was 8.43 mA, the signal-to-noise ratio improved by 9.4394 dB, and the PESQ improved by 0.7350. The presented objective and subjective results show that the proposed method achieves better noise suppression than previous methods.","PeriodicalId":45004,"journal":{"name":"Journal of Computational Methods in Sciences and Engineering","volume":" ","pages":""},"PeriodicalIF":0.5000,"publicationDate":"2023-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Computational Methods in Sciences and Engineering","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.3233/jcm-226897","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"ENGINEERING, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

Speech enhancement is essential for hearing aids. In recent years, many speech enhancement methods based on deep learning have been proven to be effective. However, these speech enhancement methods rarely consider limited hardware resources and have difficulty meeting real-time requirements, which is very important for hearing aids. To solve the above problems, we propose a method that combines beamforming and speech enhancement methods based on deep learning. Beamforming is used to filter background noise and reduce the complexity of noise. Additionally, a new filter bank used in hearing aids is adopted to reduce the complexity of the system. The system was deployed and tested in resource-constrained hearing aids. The effectiveness of the method was verified by objective experiments using standard evaluation indicators. The results showed that the power was 8.43 mA, the signal-to-noise ratio improved by 9.4394 dB, and the PESQ improved by 0.7350. The presented objective and subjective results show that the proposed method achieves better noise suppression than previous methods.

查看原文本刊更多论文

一种将波束成形与RNN相结合的助听器语音增强方法

语音增强对助听器至关重要。近年来，许多基于深度学习的语音增强方法已被证明是有效的。然而，这些语音增强方法很少考虑有限的硬件资源，难以满足实时性要求，这对助听器来说非常重要。为了解决上述问题，我们提出了一种基于深度学习的波束成形和语音增强相结合的方法。波束形成用于过滤背景噪声并降低噪声的复杂性。此外，还采用了一种用于助听器的新型滤波器组，以降低系统的复杂性。该系统在资源有限的助听器中进行了部署和测试。通过使用标准评价指标的客观实验验证了该方法的有效性。结果表明，功率为8.43mA，信噪比提高了9.4394dB，PESQ提高了0.7350。所给出的客观和主观结果表明，该方法比以前的方法具有更好的噪声抑制效果。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Computational Methods in Sciences and Engineering ENGINEERING, MULTIDISCIPLINARY-

CiteScore

0.80

自引率

0.00%

发文量

152

期刊介绍： The major goal of the Journal of Computational Methods in Sciences and Engineering (JCMSE) is the publication of new research results on computational methods in sciences and engineering. Common experience had taught us that computational methods originally developed in a given basic science, e.g. physics, can be of paramount importance to other neighboring sciences, e.g. chemistry, as well as to engineering or technology and, in turn, to society as a whole. This undoubtedly beneficial practice of interdisciplinary interactions will be continuously and systematically encouraged by the JCMSE.