Deep learning image compression with multi-channel tANS coding and hardware deployment
Deep learning-based image compression outperforms traditional methods in coding efficiency, but its computational complexity hinders real-time deployment on embedded devices. This paper proposes a heterogeneous computing system combining GPU-accelerated inference and CPU-accelerated entropy coding v...
Saved in:
| Published in: | Journal of real-time image processing Vol. 23; no. 1; p. 1 |
|---|---|
| Main Authors: | , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
Berlin/Heidelberg
Springer Berlin Heidelberg
01.01.2026
Springer Nature B.V |
| Subjects: | |
| ISSN: | 1861-8200, 1861-8219 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Deep learning-based image compression outperforms traditional methods in coding efficiency, but its computational complexity hinders real-time deployment on embedded devices. This paper proposes a heterogeneous computing system combining GPU-accelerated inference and CPU-accelerated entropy coding via lookup tables, breaking performance bottlenecks through algorithm-hardware co-design. After GPU acceleration, entropy coding becomes the dominant bottleneck (73% of runtime). To address this, we introduce three key innovations: replacing rANS with tANS encoding, converting dynamic computations into static table lookups, reducing encoding latency; a cache-friendly tANS coding scheme for the 192-channel network outputs, minimizing access latency; an out-of-range symbol encoding method, ensuring lossless and efficient compression. Experiments demonstrate that under high compression ratios, compared with traditional rANS, tANS reduces latency by 77%, with a compression ratio loss of 12.6% while still ensuring image compression quality higher than JPEG2000. |
|---|---|
| AbstractList | Deep learning-based image compression outperforms traditional methods in coding efficiency, but its computational complexity hinders real-time deployment on embedded devices. This paper proposes a heterogeneous computing system combining GPU-accelerated inference and CPU-accelerated entropy coding via lookup tables, breaking performance bottlenecks through algorithm-hardware co-design. After GPU acceleration, entropy coding becomes the dominant bottleneck (73% of runtime). To address this, we introduce three key innovations: replacing rANS with tANS encoding, converting dynamic computations into static table lookups, reducing encoding latency; a cache-friendly tANS coding scheme for the 192-channel network outputs, minimizing access latency; an out-of-range symbol encoding method, ensuring lossless and efficient compression. Experiments demonstrate that under high compression ratios, compared with traditional rANS, tANS reduces latency by 77%, with a compression ratio loss of 12.6% while still ensuring image compression quality higher than JPEG2000. |
| ArticleNumber | 1 |
| Author | Zhu, Yaohua Huang, Mingsheng Zhu, Yanghang Jiang, Jingyu Zhang, Yong Liu, Ya |
| Author_xml | – sequence: 1 givenname: Yaohua surname: Zhu fullname: Zhu, Yaohua organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences – sequence: 2 givenname: Yong surname: Zhang fullname: Zhang, Yong email: zhangyong@mail.sitp.ac.cn organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences – sequence: 3 givenname: Ya surname: Liu fullname: Liu, Ya organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences – sequence: 4 givenname: Jingyu surname: Jiang fullname: Jiang, Jingyu organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences – sequence: 5 givenname: Yanghang surname: Zhu fullname: Zhu, Yanghang organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences – sequence: 6 givenname: Mingsheng surname: Huang fullname: Huang, Mingsheng organization: Shanghai Institute of Technical Physics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Laboratory of Infrared Detection and Imaging Technology, Chinese Academy of Sciences |
| BookMark | eNp9kMlOwzAQhi1UJNrCC3CyxNlgO3GWY1VWqYIDy9Vy7EmbKrGDnarq2-MSBDcuMyPN_8_yzdDEOgsIXTJ6zSjNbwJjQqSEckEoy0tBihM0ZUXGSMFZOfmtKT1DsxC2lGZ5logp-rgF6HELytvGrnHTqTVg7breQwiNs3jfDBvc7dqhIXqjrIUWD4vn16gxR4OyBm-UN3vlARvoW3fowA7n6LRWbYCLnzxH7_d3b8tHsnp5eFouVkTznA9ElSlozUxBa1FRISqRFkrQmpscBMSTIbaYUqzOyqriKlEpcFopnRqeGJMmc3Q1zu29-9xBGOTW7byNK2XCs0KkMbKo4qNKexeCh1r2Pn7qD5JReeQnR34y8pPf_GQRTcloClFs1-D_Rv_j-gKr1HXb |
| Cites_doi | 10.34768/amcs-2023-0004 10.3390/app15052826 10.1147/rd.282.0135 10.1109/30.125072 10.1117/1.1469618 10.1109/TCSVT.2021.3101953 10.1007/s11554-024-01467-z 10.1109/T-C.1974.223784 10.1006/jvci.1997.0323 10.1007/s11554-023-01346-z 10.1109/TCOM.1972.1091168 10.1109/SMC.2017.8122605 10.3390/s25082491 10.3390/jimaging4050064 10.1016/j.bspc.2012.04.003 10.1109/JRPROC.1952.273898 10.1109/79.952804 10.1109/79.952803 |
| ContentType | Journal Article |
| Copyright | The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025. |
| Copyright_xml | – notice: The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025 Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. – notice: The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025. |
| DBID | AAYXX CITATION JQ2 |
| DOI | 10.1007/s11554-025-01795-8 |
| DatabaseName | CrossRef ProQuest Computer Science Collection |
| DatabaseTitle | CrossRef ProQuest Computer Science Collection |
| DatabaseTitleList | ProQuest Computer Science Collection |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Computer Science |
| EISSN | 1861-8219 |
| ExternalDocumentID | 10_1007_s11554_025_01795_8 |
| GroupedDBID | .VR 06D 0R~ 0VY 1N0 203 29L 2J2 2JN 2JY 2KG 2KM 2LR 2~H 30V 4.4 406 408 409 40D 40E 5VS 67Z 6NX 8TC 95- 95. 95~ 96X AAAVM AABHQ AACDK AAHNG AAIAL AAJBT AAJKR AANZL AAPKM AARTL AASML AATNV AATVU AAUYE AAWCG AAYIU AAYQN ABAKF ABBBX ABBRH ABBXA ABDBE ABDZT ABECU ABFSG ABFTV ABHLI ABHQN ABJNI ABJOX ABKCH ABKTR ABMNI ABMQK ABNWP ABQBU ABRTQ ABSXP ABTEG ABTHY ABTKH ABTMW ABWNU ABXPI ACAOD ACDTI ACGFS ACHSB ACHXU ACKNC ACMDZ ACMLO ACOKC ACOMO ACPIV ACSTC ACZOJ ADHHG ADHIR ADKNI ADKPE ADRFC ADTPH ADURQ ADYFF ADZKW AEFQL AEGAL AEGNC AEJHL AEJRE AEMSY AENEX AEOHA AEPYU AESKC AETLH AEVLU AEXYK AEZWR AFBBN AFDZB AFHIU AFLOW AFOHR AFQWF AFWTZ AFZKB AGAYW AGDGC AGMZJ AGQEE AGQMX AGRTI AGWIL AGWZB AGYKE AHAVH AHBYD AHKAY AHPBZ AHWEU AHYZX AIAKS AIGIU AIIXL AILAN AITGF AIXLP AJRNO AJZVZ ALMA_UNASSIGNED_HOLDINGS ALWAN AMKLP AMXSW AMYLF AMYQR AOCGG ARMRJ ASPBG ATHPR AVWKF AXYYD AYFIA AYJHY AZFZN B-. BA0 BSONS CS3 CSCUP DDRTE DNIVK DPUIP EBLON EBS EIOEI ESBYG FEDTE FERAY FFXSO FIGPU FNLPD FRRFC FWDCC GGCAI GGRSB GJIRD GNWQR GQ7 GQ8 GXS HF~ HG5 HG6 HLICF HMJXF HQYDN HRMNR HVGLF IJ- IKXTQ ITM IWAJR IXC IXE IZIGR IZQ I~X I~Z J-C J0Z JBSCW JCJTX JZLTJ KDC KOV LLZTM MA- N9A NPVJJ NQJWS O93 O9J OAM P9O PF0 PT4 QOS R89 R9I ROL RPX RSV S16 S1Z S27 S3B SAP SCO SDH SHX SISQX SJYHP SNE SNPRN SNX SOHCF SOJ SPISZ SRMVM SSLCW STPWE SZN T13 TSG TSK TSV TUC U2A UG4 UOJIU UTJUX UZXMN VC2 VFIZW W23 W48 WK8 YLTOR Z45 ZMTXR ~A9 -Y2 2VQ AARHV AAYTO AAYXX ABQSL ABULA ACBXY ADHKG AEBTG AEKMD AFFHD AFGCZ AFKRA AGJBK AGQPQ AHSBF AJBLW ARAPS BDATZ BENPR BGLVJ BGNMA CAG CCPQU CITATION COF EJD FINBP FSGXE H13 HCIFZ HZ~ IHE K7- M4Y N2Q NU0 O9- PHGZM PHGZT PQGLB JQ2 |
| ID | FETCH-LOGICAL-c272t-a94ecc1d80f5b055b548a50f2d7e5e861ed801aa1f69bb2a3a4e20bac4d23dd43 |
| IEDL.DBID | RSV |
| ISICitedReferencesCount | 0 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001607385600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1861-8200 |
| IngestDate | Thu Nov 06 13:46:19 EST 2025 Sat Nov 29 06:55:52 EST 2025 Wed Nov 05 01:23:39 EST 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 1 |
| Keywords | Entropy coding Deep learning image compression tANS coding Cache-friendly GPU |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c272t-a94ecc1d80f5b055b548a50f2d7e5e861ed801aa1f69bb2a3a4e20bac4d23dd43 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 |
| PQID | 3268543261 |
| PQPubID | 2044148 |
| ParticipantIDs | proquest_journals_3268543261 crossref_primary_10_1007_s11554_025_01795_8 springer_journals_10_1007_s11554_025_01795_8 |
| PublicationCentury | 2000 |
| PublicationDate | 2026-01-01 |
| PublicationDateYYYYMMDD | 2026-01-01 |
| PublicationDate_xml | – month: 01 year: 2026 text: 2026-01-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | Berlin/Heidelberg |
| PublicationPlace_xml | – name: Berlin/Heidelberg – name: Heidelberg |
| PublicationTitle | Journal of real-time image processing |
| PublicationTitleAbbrev | J Real-Time Image Proc |
| PublicationYear | 2026 |
| Publisher | Springer Berlin Heidelberg Springer Nature B.V |
| Publisher_xml | – name: Springer Berlin Heidelberg – name: Springer Nature B.V |
| References | 1795_CR7 1795_CR4 1795_CR21 MA Kabir (1795_CR9) 2018; 4 1795_CR22 B Bross (1795_CR8) 2021; 31 B Huang (1795_CR13) 2013; 8 1795_CR14 A Skodras (1795_CR6) 2001; 18 DK Lam (1795_CR11) 2024; 21 BE Usevitch (1795_CR20) 2002; 18 1795_CR3 1795_CR2 J Ballé (1795_CR1) 2016; 25 Y Zhu (1795_CR23) 2025; 25 J Pearl (1795_CR17) 1972; 20 Q Sheng (1795_CR16) 2023; 20 G Mandyam (1795_CR19) 1997; 8 N Ahmed (1795_CR18) 2006; 100 Y Zhu (1795_CR24) 2025; 15 1795_CR15 DA Huffman (1795_CR10) 1952; 40 GG Langdon (1795_CR12) 1984; 28 DS Taubman (1795_CR5) 2002; 11 |
| References_xml | – ident: 1795_CR15 doi: 10.34768/amcs-2023-0004 – volume: 15 start-page: 2826 issue: 5 year: 2025 ident: 1795_CR24 publication-title: Appl. Sci. doi: 10.3390/app15052826 – volume: 28 start-page: 135 issue: 2 year: 1984 ident: 1795_CR12 publication-title: IBM J. Res. Dev. doi: 10.1147/rd.282.0135 – ident: 1795_CR4 doi: 10.1109/30.125072 – volume: 11 start-page: 286 issue: 2 year: 2002 ident: 1795_CR5 publication-title: J. Electron. Imaging doi: 10.1117/1.1469618 – volume: 31 start-page: 3736 issue: 10 year: 2021 ident: 1795_CR8 publication-title: IEEE Trans. Circuits Syst. Video Technol. doi: 10.1109/TCSVT.2021.3101953 – ident: 1795_CR21 – volume: 25 start-page: 5706 issue: 12 year: 2016 ident: 1795_CR1 publication-title: IEEE Trans. Image Process. – ident: 1795_CR22 – volume: 21 start-page: 84 issue: 3 year: 2024 ident: 1795_CR11 publication-title: J. Real-Time Image Proc. doi: 10.1007/s11554-024-01467-z – volume: 100 start-page: 90 issue: 1 year: 2006 ident: 1795_CR18 publication-title: IEEE Trans. Comput. doi: 10.1109/T-C.1974.223784 – volume: 8 start-page: 21 issue: 1 year: 1997 ident: 1795_CR19 publication-title: J. Vis. Commun. Image Represent. doi: 10.1006/jvci.1997.0323 – ident: 1795_CR14 – volume: 20 start-page: 90 issue: 5 year: 2023 ident: 1795_CR16 publication-title: J. Real-Time Image Proc. doi: 10.1007/s11554-023-01346-z – volume: 20 start-page: 411 issue: 3 year: 1972 ident: 1795_CR17 publication-title: IEEE Trans. Commun. doi: 10.1109/TCOM.1972.1091168 – ident: 1795_CR3 – ident: 1795_CR7 doi: 10.1109/SMC.2017.8122605 – volume: 25 start-page: 2491 issue: 8 year: 2025 ident: 1795_CR23 publication-title: Sensors doi: 10.3390/s25082491 – ident: 1795_CR2 – volume: 4 start-page: 64 issue: 5 year: 2018 ident: 1795_CR9 publication-title: J. Imaging doi: 10.3390/jimaging4050064 – volume: 8 start-page: 59 issue: 1 year: 2013 ident: 1795_CR13 publication-title: Biomed. Signal Process. Control doi: 10.1016/j.bspc.2012.04.003 – volume: 40 start-page: 1098 issue: 9 year: 1952 ident: 1795_CR10 publication-title: Proc. IRE doi: 10.1109/JRPROC.1952.273898 – volume: 18 start-page: 36 issue: 5 year: 2001 ident: 1795_CR6 publication-title: IEEE Signal Process. Mag. doi: 10.1109/79.952804 – volume: 18 start-page: 22 issue: 5 year: 2002 ident: 1795_CR20 publication-title: IEEE Signal Process. Mag. doi: 10.1109/79.952803 |
| SSID | ssj0067635 |
| Score | 2.363297 |
| Snippet | Deep learning-based image compression outperforms traditional methods in coding efficiency, but its computational complexity hinders real-time deployment on... |
| SourceID | proquest crossref springer |
| SourceType | Aggregation Database Index Database Publisher |
| StartPage | 1 |
| SubjectTerms | Algorithms Co-design Coding Compression ratio Computer Graphics Computer Science Deep learning Efficiency Embedded systems Entropy Graphics processing units Hardware Image compression Image Processing and Computer Vision Image quality Lookup tables Multimedia Information Systems Neural networks Pattern Recognition Real time Signal,Image and Speech Processing Wavelet transforms |
| Title | Deep learning image compression with multi-channel tANS coding and hardware deployment |
| URI | https://link.springer.com/article/10.1007/s11554-025-01795-8 https://www.proquest.com/docview/3268543261 |
| Volume | 23 |
| WOSCitedRecordID | wos001607385600001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVAVX databaseName: SpringerLink customDbUrl: eissn: 1861-8219 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0067635 issn: 1861-8200 databaseCode: RSV dateStart: 20060301 isFulltext: true titleUrlDefault: https://link.springer.com/search?facet-content-type=%22Journal%22 providerName: Springer Nature |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV07T8MwELagMLBQnqJQkAc2sJT4kTpjBVQMqEIqVN0iO76gShCqNsDfx5eHIhAMsGSxdYrO53v57jtCzqUKMuEjCwbYnitBBkxnacaUdjEHEYu0AnG9G4zHejaL7-umsFVT7d48SZaaum12Q9PHcPwqSpFiep1sKMTRwxh9Mm30b4QQaxhm6Shk3r4FdavMzzS-mqPWx_z2LFpam1H3f_-5Q7Zr75IOK3HYJWuQ75FuM7mB1hd5n0yvARa0nhjxROcvXqtQrC6vqmJziulZWhYbMmwNzuGZFsPxxO9BW0dN7ii2a32YJVAHODQYs4wH5HF083B1y-oJCyzlA14wE0t_hKHTQaZsoJT18Yvxp8fdABR4BoJfCo0Jsyi2lhthJPDAmlQ6LpyT4pB08tccjgjVEEbW6hAQnNRqobVB2yeyzKSeeNQjFw2jk0UFpJG0kMnIssSzLClZluge6TdnkdSXapV4T1Mr6b9hj1w2vG-Xf6d2_LftJ2SL-9C0SrT0SadYvsEp2Uzfi_lqeVYK2ydvOM6X |
| linkProvider | Springer Nature |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV1LS8NAEB60CnqxPrFadQ_edCHZbNLNsailYg1Ca-ktbJKJFDSWNurfdycPiqIHveSyyxC-nZ3XzgPgXLpW6hjPgiOV50qUFldpnHJXJb5Ax3fisonroBMEajLxH6qisEWd7V4_SRaSelnsRqqP0_hV4iKXq1VYk8LIYPLRh-Na_nrUYo3cLOXZ3Og3qyqV-ZnGV3W0tDG_PYsW2qbX_N9_bsNWZV2ybskOO7CC2S4068kNrLrIezC-RpyxamLEE5u-GKnCKLu8zIrNGIVnWZFsyKk0OMNnlneDodlDuo7pLGFUrvWh58gSpKHBFGXch8fezeiqz6sJCzwWHZFz7UtzhHairNSNLNeNjP-izemJpIMuGgDRLNla26nnR5HQjpYorEjHMhFOkkjnABrZa4aHwBTaXhQpG6k5aaQcpTTpPidNdWyIey24qIEOZ2UjjXDZMpkgCw1kYQFZqFrQrs8irC7VIjSWpnKl-dotuKyxXy7_Tu3ob9vPYKM_uh-Eg9vg7hg2hXFTy6BLGxr5_A1PYD1-z6eL-WnBeJ8RutFy |
| linkToPdf | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LS8NAEB60inixPrFadQ_edGkem3RzLNaiWEqhWnoLm-xEChpLG_Xvu5MHVdGDeMlllyHMzO48dr4ZgHPhWYlrIguOBM8VKCwukzjhntSBg27gxkUT1357MJCTSTD8hOLPq92rJ8kC00BdmtKsNdNJawl8IzPIaRQraZTH5SqsCYpuKF4fjau72Kd2axRySd_mxtZZJWzmZxpfTdPS3_z2RJpbnl79__-8DVul18k6hZrswAqmu1CvJjqw8oDvwbiLOGPlJIlHNn02tw2jqvOiWjZllLZleREiJ8hwik8s6wxGZg_ZQKZSzQjG9a7myDTSMGHKPu7DQ-_6_uqGl5MXeOy0nYyrQBjR2lpaiRdZnheZuEYZqTq6jR4aZqJZspWyEz-IIke5SqBjRSoW2nG1Fu4B1NKXFA-BSbT9KJI2UtPSSLpSKrKJbpKo2BD3G3BRMT2cFQ02wmUrZWJZaFgW5iwLZQOalVzC8rAtQuOBSk-Yr92Ay0oOy-XfqR39bfsZbAy7vbB_O7g7hk3HRK9FLqYJtWz-iiewHr9l08X8NNfBD9Rd2lY |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Deep+learning+image+compression+with+multi-channel+tANS+coding+and+hardware+deployment&rft.jtitle=Journal+of+real-time+image+processing&rft.au=Zhu%2C+Yaohua&rft.au=Zhang%2C+Yong&rft.au=Liu%2C+Ya&rft.au=Jiang%2C+Jingyu&rft.date=2026-01-01&rft.pub=Springer+Berlin+Heidelberg&rft.issn=1861-8200&rft.eissn=1861-8219&rft.volume=23&rft.issue=1&rft_id=info:doi/10.1007%2Fs11554-025-01795-8&rft.externalDocID=10_1007_s11554_025_01795_8 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1861-8200&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1861-8200&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1861-8200&client=summon |