PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement
Low-complexity speech enhancement on mobile phones is crucial in the era of 5G. Thus, focusing on handheld mobile phone communication scenario, based on power level difference (PLD) algorithm and lightweight U-Net, we propose PLD-guided lightweight deep network (PLDNet), an extremely lightweight dua...
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Artikel |
| Sprache: | eng |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | |
|---|---|
| container_issue | |
| container_start_page | |
| container_title | |
| container_volume | |
| creator | Zhou, Nan Jiang, Youhai Tan, Jialin Qi, Chongmin |
| description | Low-complexity speech enhancement on mobile phones is crucial in the era of
5G. Thus, focusing on handheld mobile phone communication scenario, based on
power level difference (PLD) algorithm and lightweight U-Net, we propose
PLD-guided lightweight deep network (PLDNet), an extremely lightweight
dual-microphone speech enhancement method that integrates the guidance of
signal processing algorithm and lightweight attention-augmented U-Net. For the
guidance information, we employ PLD algorithm to pre-process dual-microphone
spectrum, and feed the output into subsequent deep neural network, which
utilizes a lightweight U-Net with our proposed gated convolution augmented
frequency attention (GCAFA) module to extract desired clean speech.
Experimental results demonstrate that our proposed method achieves competitive
performance with recent top-performing models while reducing computational cost
by over 90%, highlighting the potential for low-complexity speech enhancement
on mobile phones. |
| doi_str_mv | 10.48550/arxiv.2406.03899 |
| format | Article |
| fullrecord | <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2406_03899</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2406_03899</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-db5ec306e925584e40f613f85c19b6b96f5d249704f745c61e1d27ee0a62414c3</originalsourceid><addsrcrecordid>eNotkL1OwzAcxL0woMIDMOEXSLAT24nZSlNapPAh0T1y7L-JRWpHbkopT09aWO633J10h9ANJSkrOSd3Kn67rzRjRKQkL6W8RD9vdfUC4z2emKz2zoDBtfvoxgOcFFcAA54MhxA_8UMIu3EytEe8tNZpB37E83Gc4ILHNkS8Vt500Btc7VWfPDsdw9AFD_h9ANAdXvpOeQ3bKXKFLqzqd3D9zxnaPC43i3VSv66eFvM6UaKQiWk56JwIkBnnJQNGrKC5LbmmshWtFJabjMmCMFswrgUFarICgCiRMcp0PkO3f7Xn8c0Q3VbFY3M6oTmfkP8CszZXjw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement</title><source>arXiv.org</source><creator>Zhou, Nan ; Jiang, Youhai ; Tan, Jialin ; Qi, Chongmin</creator><creatorcontrib>Zhou, Nan ; Jiang, Youhai ; Tan, Jialin ; Qi, Chongmin</creatorcontrib><description>Low-complexity speech enhancement on mobile phones is crucial in the era of
5G. Thus, focusing on handheld mobile phone communication scenario, based on
power level difference (PLD) algorithm and lightweight U-Net, we propose
PLD-guided lightweight deep network (PLDNet), an extremely lightweight
dual-microphone speech enhancement method that integrates the guidance of
signal processing algorithm and lightweight attention-augmented U-Net. For the
guidance information, we employ PLD algorithm to pre-process dual-microphone
spectrum, and feed the output into subsequent deep neural network, which
utilizes a lightweight U-Net with our proposed gated convolution augmented
frequency attention (GCAFA) module to extract desired clean speech.
Experimental results demonstrate that our proposed method achieves competitive
performance with recent top-performing models while reducing computational cost
by over 90%, highlighting the potential for low-complexity speech enhancement
on mobile phones.</description><identifier>DOI: 10.48550/arxiv.2406.03899</identifier><language>eng</language><creationdate>2024-06</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2406.03899$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2406.03899$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Zhou, Nan</creatorcontrib><creatorcontrib>Jiang, Youhai</creatorcontrib><creatorcontrib>Tan, Jialin</creatorcontrib><creatorcontrib>Qi, Chongmin</creatorcontrib><title>PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement</title><description>Low-complexity speech enhancement on mobile phones is crucial in the era of
5G. Thus, focusing on handheld mobile phone communication scenario, based on
power level difference (PLD) algorithm and lightweight U-Net, we propose
PLD-guided lightweight deep network (PLDNet), an extremely lightweight
dual-microphone speech enhancement method that integrates the guidance of
signal processing algorithm and lightweight attention-augmented U-Net. For the
guidance information, we employ PLD algorithm to pre-process dual-microphone
spectrum, and feed the output into subsequent deep neural network, which
utilizes a lightweight U-Net with our proposed gated convolution augmented
frequency attention (GCAFA) module to extract desired clean speech.
Experimental results demonstrate that our proposed method achieves competitive
performance with recent top-performing models while reducing computational cost
by over 90%, highlighting the potential for low-complexity speech enhancement
on mobile phones.</description><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotkL1OwzAcxL0woMIDMOEXSLAT24nZSlNapPAh0T1y7L-JRWpHbkopT09aWO633J10h9ANJSkrOSd3Kn67rzRjRKQkL6W8RD9vdfUC4z2emKz2zoDBtfvoxgOcFFcAA54MhxA_8UMIu3EytEe8tNZpB37E83Gc4ILHNkS8Vt500Btc7VWfPDsdw9AFD_h9ANAdXvpOeQ3bKXKFLqzqd3D9zxnaPC43i3VSv66eFvM6UaKQiWk56JwIkBnnJQNGrKC5LbmmshWtFJabjMmCMFswrgUFarICgCiRMcp0PkO3f7Xn8c0Q3VbFY3M6oTmfkP8CszZXjw</recordid><startdate>20240606</startdate><enddate>20240606</enddate><creator>Zhou, Nan</creator><creator>Jiang, Youhai</creator><creator>Tan, Jialin</creator><creator>Qi, Chongmin</creator><scope>GOX</scope></search><sort><creationdate>20240606</creationdate><title>PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement</title><author>Zhou, Nan ; Jiang, Youhai ; Tan, Jialin ; Qi, Chongmin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-db5ec306e925584e40f613f85c19b6b96f5d249704f745c61e1d27ee0a62414c3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><toplevel>online_resources</toplevel><creatorcontrib>Zhou, Nan</creatorcontrib><creatorcontrib>Jiang, Youhai</creatorcontrib><creatorcontrib>Tan, Jialin</creatorcontrib><creatorcontrib>Qi, Chongmin</creatorcontrib><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zhou, Nan</au><au>Jiang, Youhai</au><au>Tan, Jialin</au><au>Qi, Chongmin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement</atitle><date>2024-06-06</date><risdate>2024</risdate><abstract>Low-complexity speech enhancement on mobile phones is crucial in the era of
5G. Thus, focusing on handheld mobile phone communication scenario, based on
power level difference (PLD) algorithm and lightweight U-Net, we propose
PLD-guided lightweight deep network (PLDNet), an extremely lightweight
dual-microphone speech enhancement method that integrates the guidance of
signal processing algorithm and lightweight attention-augmented U-Net. For the
guidance information, we employ PLD algorithm to pre-process dual-microphone
spectrum, and feed the output into subsequent deep neural network, which
utilizes a lightweight U-Net with our proposed gated convolution augmented
frequency attention (GCAFA) module to extract desired clean speech.
Experimental results demonstrate that our proposed method achieves competitive
performance with recent top-performing models while reducing computational cost
by over 90%, highlighting the potential for low-complexity speech enhancement
on mobile phones.</abstract><doi>10.48550/arxiv.2406.03899</doi><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | DOI: 10.48550/arxiv.2406.03899 |
| ispartof | |
| issn | |
| language | eng |
| recordid | cdi_arxiv_primary_2406_03899 |
| source | arXiv.org |
| title | PLDNet: PLD-Guided Lightweight Deep Network Boosted by Efficient Attention for Handheld Dual-Microphone Speech Enhancement |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-09T14%3A54%3A52IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=PLDNet:%20PLD-Guided%20Lightweight%20Deep%20Network%20Boosted%20by%20Efficient%20Attention%20for%20Handheld%20Dual-Microphone%20Speech%20Enhancement&rft.au=Zhou,%20Nan&rft.date=2024-06-06&rft_id=info:doi/10.48550/arxiv.2406.03899&rft_dat=%3Carxiv_GOX%3E2406_03899%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |