两点神经元用于有效的多模态语音增强

2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW) Pub Date : 2023-06-04 DOI:10.1109/ICASSPW59220.2023.10193457

M. Raza, Khubaib Ahmed, Junaid Muzaffar, Ahsan Adeel

{"title":"两点神经元用于有效的多模态语音增强","authors":"M. Raza, Khubaib Ahmed, Junaid Muzaffar, Ahsan Adeel","doi":"10.1109/ICASSPW59220.2023.10193457","DOIUrl":null,"url":null,"abstract":"Here we present a two-point neuron-inspired deep convolutional net (DCN) with 18 convolutional layers for multimodal speech enhancement (MM-SE) and compare it against conventional point neuron-inspired DCN in terms of Perceptual Evaluation of Speech Quality (PESQ) and Short-Time Objective Intelligibility (STOI). We show that the two-point neuron-driven DCN performs comparably to point-neurons driven DCN by using only ≈0.2% neurons at any time during training.","PeriodicalId":158726,"journal":{"name":"2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-06-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Two-Point Neurons for Efficient Multimodal Speech Enhancement\",\"authors\":\"M. Raza, Khubaib Ahmed, Junaid Muzaffar, Ahsan Adeel\",\"doi\":\"10.1109/ICASSPW59220.2023.10193457\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Here we present a two-point neuron-inspired deep convolutional net (DCN) with 18 convolutional layers for multimodal speech enhancement (MM-SE) and compare it against conventional point neuron-inspired DCN in terms of Perceptual Evaluation of Speech Quality (PESQ) and Short-Time Objective Intelligibility (STOI). We show that the two-point neuron-driven DCN performs comparably to point-neurons driven DCN by using only ≈0.2% neurons at any time during training.\",\"PeriodicalId\":158726,\"journal\":{\"name\":\"2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)\",\"volume\":\"56 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2023-06-04\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICASSPW59220.2023.10193457\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICASSPW59220.2023.10193457","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 0

摘要

本文提出了一个具有18个卷积层的多点神经元启发深度卷积网络(DCN)，用于多模态语音增强(MM-SE)，并在语音质量感知评价(PESQ)和短时客观可理解性(STOI)方面与传统点神经元启发的DCN进行了比较。我们证明两点神经元驱动的DCN与点神经元驱动的DCN相比，在训练过程中任何时间只使用≈0.2%的神经元。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Two-Point Neurons for Efficient Multimodal Speech Enhancement

Here we present a two-point neuron-inspired deep convolutional net (DCN) with 18 convolutional layers for multimodal speech enhancement (MM-SE) and compare it against conventional point neuron-inspired DCN in terms of Perceptual Evaluation of Speech Quality (PESQ) and Short-Time Objective Intelligibility (STOI). We show that the two-point neuron-driven DCN performs comparably to point-neurons driven DCN by using only ≈0.2% neurons at any time during training.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW)

自引率

0.00%

发文量