Multi-rate deep semantic image compression with quantized modulated autoencoder

Recently, deep learning has demonstrated impressive performance in image compression. Methods, that achieve and even outperform conventional codecs performances, are continually emerging. However, most of them need to train and deploy separate networks for rate adaptation. This is impractical and ex...

Full description

Saved in:
Bibliographic Details
Published in:IEEE International Workshop on Multimedia Signal Processing pp. 1 - 6
Main Author: Sebai, Dorsaf
Format: Conference Proceeding
Language:English
Published: IEEE 06.10.2021
Subjects:
Computational modeling
Conditional autoencoder
Deep compression
Image coding
Memory management
Quantization (signal)
Quantized autoencoder
Semantic analysis
Semantics
Training
Variable-rate compression
Visualization
ISSN:2473-3628
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract Recently, deep learning has demonstrated impressive performance in image compression. Methods, that achieve and even outperform conventional codecs performances, are continually emerging. However, most of them need to train and deploy separate networks for rate adaptation. This is impractical and extensive in terms of memory cost and power consumption, especially for broad bitrate ranges. Further, methods that consider the semantic-important structure of the image are extremely sparse. This leads to non-optimized bit allocation for the eye-catching foreground details, that have to be preserved for the almost all computer vision applications. Towards this end, we establish an end-to-end multi-rate deep semantic image compression with quantized conditional autoencoder. It includes two neural networks for the semantic analysis and image compression, respectively. The semantic analysis network extracts the essential semantic regions of the input image, and calculates the Semantic-Important Structural SIMilarity (SI-SSIM) index for each of them. The compression network is then trained to optimize a multi-loss function based on SI-SSIM and conditioned on the activation bitwidths. Performances of our model are evaluated on the JPEG AI dataset for objective and perceptual quality metrics. Obtained results show that our method yields higher performances over JPEG, JPEG 2000 and HEVC intra baselines and competitive performances with VVC intra.
AbstractList Recently, deep learning has demonstrated impressive performance in image compression. Methods, that achieve and even outperform conventional codecs performances, are continually emerging. However, most of them need to train and deploy separate networks for rate adaptation. This is impractical and extensive in terms of memory cost and power consumption, especially for broad bitrate ranges. Further, methods that consider the semantic-important structure of the image are extremely sparse. This leads to non-optimized bit allocation for the eye-catching foreground details, that have to be preserved for the almost all computer vision applications. Towards this end, we establish an end-to-end multi-rate deep semantic image compression with quantized conditional autoencoder. It includes two neural networks for the semantic analysis and image compression, respectively. The semantic analysis network extracts the essential semantic regions of the input image, and calculates the Semantic-Important Structural SIMilarity (SI-SSIM) index for each of them. The compression network is then trained to optimize a multi-loss function based on SI-SSIM and conditioned on the activation bitwidths. Performances of our model are evaluated on the JPEG AI dataset for objective and perceptual quality metrics. Obtained results show that our method yields higher performances over JPEG, JPEG 2000 and HEVC intra baselines and competitive performances with VVC intra.
Author Sebai, Dorsaf
Author_xml – sequence: 1
  givenname: Dorsaf
  surname: Sebai
  fullname: Sebai, Dorsaf
  organization: National School of Computer Science (ENSI),Cristal Laboratory,Manouba,Tunisia
BookMark eNotkMtKxDAYhaMoOI59AkHyAq25X5YyeIMpI6jrIU3-amTa1KZF9OkdcVZncfgOH-ccnfSpB4SuKKkoJfa6rp-fJCdUV4wwWlnNuZTkCBVWG6qUFJwZo4_RggnNS66YOUNFzh-EEKqYMIws0Kaed1MsRzcBDgADztC5fooex869AfapG0bIOaYef8XpHX_Of_UPBNylMO_2XMBunhL0PgUYL9Bp63YZikMu0evd7cvqoVxv7h9XN-syUmqm0gjWMO-Vo7KhLVfBaBGksyCCC42XbeuNFbY1zjstmBaqCcE5arWVvNWBL9Hl_24EgO0w7m3H7-3hAv4L3zNUcg
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/MMSP53017.2021.9733550
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISBN 9781665432887
1665432888
EISSN 2473-3628
EndPage 6
ExternalDocumentID 9733550
Genre orig-research
GroupedDBID 6IE
6IF
6IK
6IL
6IN
AAJGR
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IPLJI
OCL
RIE
RIL
RNS
ID FETCH-LOGICAL-i118t-842b2cc6a15b1f36d874d5a9e4dadbc5ffc8949f8aca742746bddaa197953f7d3
IEDL.DBID RIE
IngestDate Wed Aug 27 02:24:28 EDT 2025
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i118t-842b2cc6a15b1f36d874d5a9e4dadbc5ffc8949f8aca742746bddaa197953f7d3
PageCount 6
ParticipantIDs ieee_primary_9733550
PublicationCentury 2000
PublicationDate 2021-Oct.-6
PublicationDateYYYYMMDD 2021-10-06
PublicationDate_xml – month: 10
  year: 2021
  text: 2021-Oct.-6
  day: 06
PublicationDecade 2020
PublicationTitle IEEE International Workshop on Multimedia Signal Processing
PublicationTitleAbbrev MMSP
PublicationYear 2021
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0001624820
Score 1.7865824
Snippet Recently, deep learning has demonstrated impressive performance in image compression. Methods, that achieve and even outperform conventional codecs...
SourceID ieee
SourceType Publisher
StartPage 1
SubjectTerms Computational modeling
Conditional autoencoder
Deep compression
Image coding
Memory management
Quantization (signal)
Quantized autoencoder
Semantic analysis
Semantics
Training
Variable-rate compression
Visualization
Title Multi-rate deep semantic image compression with quantized modulated autoencoder
URI https://ieeexplore.ieee.org/document/9733550
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELZKxQALjxbxlgdG0jaJHdszomKgpRIgdav8OEsZmpS2YeDXc06jFiQWNitWFOls57s73_cdIXfoIahwXYOem9MRs4kIR8pEAMwb72yqdK2u_yzGYzmdqkmL3G-5MABQF59BLwzru3xX2iqkyvpKpAiPGKDvCZFtuFq7fEqWMESzhgQcD1R_NHqdcNy_AqPAJO41L__qolKDyPDof58_Jt0dG49OtjhzQlpQnJLDH0KCHfJS82ijIPtAHcCCrmCOJsstzef4w6ChcHxT8FrQkHmlH1WY_gJH56ULHbxwpKt1GWQtHSy75H34-PbwFDWtEqIcI4R1JFliEmszHXMT-zRzUjDHtQLmtDOWe2-lYspLbTUGw4JlxjmtYyUUT71w6RlpF2UB54TGHl2uoHMvEfnBcj3gBmHMyiw23nN5QTrBNLPFRg1j1ljl8u_HV-QgWL8uf8uuSXu9rOCG7NvPdb5a3tZL-A2aMJ6G
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07T8MwELaqggQsPFrEGw-MpG0SO7ZnRFVEWypRpG6VH2cpQ5PSJgz8euw0oiCxsFmJIllnx9_d-b7vELpzHoLw1zXOczMyIDpi_pdSAQCxyhodC1mp6w_ZeMxnMzFpoPtvLgwAVMVn0PHD6i7f5Lr0qbKuYLGDRxeg71BCot6GrbXNqCQRcXhW04DDnuiORq8T6nYwc3FgFHbqz3_1UalgpH_4vwkcofaWj4cn30hzjBqQnaCDH1KCLfRSMWkDL_yADcASr2HhjJZqnC7ckYF96fim5DXDPveK30v_-hMMXuTG9_ByI1kWuRe2NLBqo7f-4_RhENTNEoLUxQhFwEmkIq0TGVIV2jgxnBFDpQBipFGaWqu5IMJyqaULhxlJlDFShoIJGltm4lPUzPIMzhAOrXO6vNI9d9gPmsoeVQ7INE9CZS3l56jlTTNfbvQw5rVVLv5-fIv2BtPRcD58Gj9fon2_ElUxXHKFmsWqhGu0qz-KdL26qZbzCwc6oc0
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%3Abook&rft.genre=proceeding&rft.title=IEEE+International+Workshop+on+Multimedia+Signal+Processing&rft.atitle=Multi-rate+deep+semantic+image+compression+with+quantized+modulated+autoencoder&rft.au=Sebai%2C+Dorsaf&rft.date=2021-10-06&rft.pub=IEEE&rft.eissn=2473-3628&rft.spage=1&rft.epage=6&rft_id=info:doi/10.1109%2FMMSP53017.2021.9733550&rft.externalDocID=9733550