Loading…

JPEG Inspired Deep Learning

Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve the performance of deep learning (DL). Inspired b...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-10
Main Authors: Salamah, Ahmed H, Zheng, Kaixiang, Liu, Yiwen, En-Hui, Yang
Format: Article
Language:English
Subjects:
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Salamah, Ahmed H
Zheng, Kaixiang
Liu, Yiwen
En-Hui, Yang
description Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve the performance of deep learning (DL). Inspired by this, we propose JPEG-DL, a novel DL framework that prepends any underlying DNN architecture with a trainable JPEG compression layer. To make the quantization operation in JPEG compression trainable, a new differentiable soft quantizer is employed at the JPEG layer, and then the quantization operation and underlying DNN are jointly trained. Extensive experiments show that in comparison with the standard DL, JPEG-DL delivers significant accuracy improvements across various datasets and model architectures while enhancing robustness against adversarial attacks. Particularly, on some fine-grained image classification datasets, JPEG-DL can increase prediction accuracy by as much as 20.9%. Our code is available on https://github.com/JpegInspiredDl/JPEG-Inspired-DL.git.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_3115207116</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3115207116</sourcerecordid><originalsourceid>FETCH-proquest_journals_31152071163</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQ9gpwdVfwzCsuyCxKTVFwSU0tUPBJTSzKy8xL52FgTUvMKU7lhdLcDMpuriHOHroFRfmFpanFJfFZ-aVFeUCpeGNDQ1MjA3NDQzNj4lQBAJZkKkg</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3115207116</pqid></control><display><type>article</type><title>JPEG Inspired Deep Learning</title><source>Publicly Available Content Database</source><creator>Salamah, Ahmed H ; Zheng, Kaixiang ; Liu, Yiwen ; En-Hui, Yang</creator><creatorcontrib>Salamah, Ahmed H ; Zheng, Kaixiang ; Liu, Yiwen ; En-Hui, Yang</creatorcontrib><description>Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve the performance of deep learning (DL). Inspired by this, we propose JPEG-DL, a novel DL framework that prepends any underlying DNN architecture with a trainable JPEG compression layer. To make the quantization operation in JPEG compression trainable, a new differentiable soft quantizer is employed at the JPEG layer, and then the quantization operation and underlying DNN are jointly trained. Extensive experiments show that in comparison with the standard DL, JPEG-DL delivers significant accuracy improvements across various datasets and model architectures while enhancing robustness against adversarial attacks. Particularly, on some fine-grained image classification datasets, JPEG-DL can increase prediction accuracy by as much as 20.9%. Our code is available on https://github.com/JpegInspiredDl/JPEG-Inspired-DL.git.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Artificial neural networks ; Datasets ; Deep learning ; Image classification ; Image compression ; Image enhancement ; Machine learning ; Performance enhancement</subject><ispartof>arXiv.org, 2024-10</ispartof><rights>2024. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://www.proquest.com/docview/3115207116?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>776,780,25732,36991,44569</link.rule.ids></links><search><creatorcontrib>Salamah, Ahmed H</creatorcontrib><creatorcontrib>Zheng, Kaixiang</creatorcontrib><creatorcontrib>Liu, Yiwen</creatorcontrib><creatorcontrib>En-Hui, Yang</creatorcontrib><title>JPEG Inspired Deep Learning</title><title>arXiv.org</title><description>Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve the performance of deep learning (DL). Inspired by this, we propose JPEG-DL, a novel DL framework that prepends any underlying DNN architecture with a trainable JPEG compression layer. To make the quantization operation in JPEG compression trainable, a new differentiable soft quantizer is employed at the JPEG layer, and then the quantization operation and underlying DNN are jointly trained. Extensive experiments show that in comparison with the standard DL, JPEG-DL delivers significant accuracy improvements across various datasets and model architectures while enhancing robustness against adversarial attacks. Particularly, on some fine-grained image classification datasets, JPEG-DL can increase prediction accuracy by as much as 20.9%. Our code is available on https://github.com/JpegInspiredDl/JPEG-Inspired-DL.git.</description><subject>Artificial neural networks</subject><subject>Datasets</subject><subject>Deep learning</subject><subject>Image classification</subject><subject>Image compression</subject><subject>Image enhancement</subject><subject>Machine learning</subject><subject>Performance enhancement</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQ9gpwdVfwzCsuyCxKTVFwSU0tUPBJTSzKy8xL52FgTUvMKU7lhdLcDMpuriHOHroFRfmFpanFJfFZ-aVFeUCpeGNDQ1MjA3NDQzNj4lQBAJZkKkg</recordid><startdate>20241009</startdate><enddate>20241009</enddate><creator>Salamah, Ahmed H</creator><creator>Zheng, Kaixiang</creator><creator>Liu, Yiwen</creator><creator>En-Hui, Yang</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20241009</creationdate><title>JPEG Inspired Deep Learning</title><author>Salamah, Ahmed H ; Zheng, Kaixiang ; Liu, Yiwen ; En-Hui, Yang</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_31152071163</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Artificial neural networks</topic><topic>Datasets</topic><topic>Deep learning</topic><topic>Image classification</topic><topic>Image compression</topic><topic>Image enhancement</topic><topic>Machine learning</topic><topic>Performance enhancement</topic><toplevel>online_resources</toplevel><creatorcontrib>Salamah, Ahmed H</creatorcontrib><creatorcontrib>Zheng, Kaixiang</creatorcontrib><creatorcontrib>Liu, Yiwen</creatorcontrib><creatorcontrib>En-Hui, Yang</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Databases</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Salamah, Ahmed H</au><au>Zheng, Kaixiang</au><au>Liu, Yiwen</au><au>En-Hui, Yang</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>JPEG Inspired Deep Learning</atitle><jtitle>arXiv.org</jtitle><date>2024-10-09</date><risdate>2024</risdate><eissn>2331-8422</eissn><abstract>Although it is traditionally believed that lossy image compression, such as JPEG compression, has a negative impact on the performance of deep neural networks (DNNs), it is shown by recent works that well-crafted JPEG compression can actually improve the performance of deep learning (DL). Inspired by this, we propose JPEG-DL, a novel DL framework that prepends any underlying DNN architecture with a trainable JPEG compression layer. To make the quantization operation in JPEG compression trainable, a new differentiable soft quantizer is employed at the JPEG layer, and then the quantization operation and underlying DNN are jointly trained. Extensive experiments show that in comparison with the standard DL, JPEG-DL delivers significant accuracy improvements across various datasets and model architectures while enhancing robustness against adversarial attacks. Particularly, on some fine-grained image classification datasets, JPEG-DL can increase prediction accuracy by as much as 20.9%. Our code is available on https://github.com/JpegInspiredDl/JPEG-Inspired-DL.git.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2024-10
issn 2331-8422
language eng
recordid cdi_proquest_journals_3115207116
source Publicly Available Content Database
subjects Artificial neural networks
Datasets
Deep learning
Image classification
Image compression
Image enhancement
Machine learning
Performance enhancement
title JPEG Inspired Deep Learning
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-24T22%3A14%3A04IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=JPEG%20Inspired%20Deep%20Learning&rft.jtitle=arXiv.org&rft.au=Salamah,%20Ahmed%20H&rft.date=2024-10-09&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E3115207116%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_31152071163%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=3115207116&rft_id=info:pmid/&rfr_iscdi=true