Learned distributed image compression with decoder side information

With the rapid development of digital communication and the widespread use of the Internet of Things, multi-view image compression has attracted increasing attention as a fundamental technology for image data communication. Multi-view image compression aims to improve compression efficiency by lever...

Full description

Saved in:
Bibliographic Details
Main Authors: Yankai Yin, Zhe Sun, Peiying Ruan, Ruidong Li, Feng Duan
Format: Article
Language:English
Published: KeAi Communications Co., Ltd. 2025-04-01
Series:Digital Communications and Networks
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2352864824000683
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849323382705750016
author Yankai Yin
Zhe Sun
Peiying Ruan
Ruidong Li
Feng Duan
author_facet Yankai Yin
Zhe Sun
Peiying Ruan
Ruidong Li
Feng Duan
author_sort Yankai Yin
collection DOAJ
description With the rapid development of digital communication and the widespread use of the Internet of Things, multi-view image compression has attracted increasing attention as a fundamental technology for image data communication. Multi-view image compression aims to improve compression efficiency by leveraging correlations between images. However, the requirement of synchronization and inter-image communication at the encoder side poses significant challenges, especially for constrained devices. In this study, we introduce a novel distributed image compression model based on the attention mechanism to address the challenges associated with the availability of side information only during decoding. Our model integrates an encoder network, a quantization module, and a decoder network, to ensure both high compression performance and high-quality image reconstruction. The encoder uses a deep Convolutional Neural Network (CNN) to extract high-level features from the input image, which then pass through the quantization module for further compression before undergoing lossless entropy coding. The decoder of our model consists of three main components that allow us to fully exploit the information within and between images on the decoder side. Specifically, we first introduce a channel-spatial attention module to capture and refine information within individual image feature maps. Second, we employ a semi-coupled convolution module to extract both shared and specific information in images. Finally, a cross-attention module is employed to fuse mutual information extracted from side information. The effectiveness of our model is validated on various datasets, including KITTI Stereo and Cityscapes. The results highlight the superior compression capabilities of our method, surpassing state-of-the-art techniques.
format Article
id doaj-art-42e2ba73ba72442dbc424dd27f42b485
institution Kabale University
issn 2352-8648
language English
publishDate 2025-04-01
publisher KeAi Communications Co., Ltd.
record_format Article
series Digital Communications and Networks
spelling doaj-art-42e2ba73ba72442dbc424dd27f42b4852025-08-20T03:49:03ZengKeAi Communications Co., Ltd.Digital Communications and Networks2352-86482025-04-0111234935810.1016/j.dcan.2024.06.001Learned distributed image compression with decoder side informationYankai Yin0Zhe Sun1Peiying Ruan2Ruidong Li3Feng Duan4Tianjin Key Laboratory of Interventional Brain-Computer Interface and Intelligent Rehabilitation, Nankai University, Tianjin 300350, China; Institute of Natural Sciences, Kanazawa University, Ishikawa 9201164, JapanFaculty of Health Data Science and Graduate School of Medicine, Juntendo University, Chiba 2790013, JapanNVIDIA AI Technology Center, NVIDIA Japan, Tokyo 1070052, JapanInstitute of Natural Sciences, Kanazawa University, Ishikawa 9201164, Japan; Corresponding authors.Tianjin Key Laboratory of Interventional Brain-Computer Interface and Intelligent Rehabilitation, Nankai University, Tianjin 300350, China; Corresponding authors.With the rapid development of digital communication and the widespread use of the Internet of Things, multi-view image compression has attracted increasing attention as a fundamental technology for image data communication. Multi-view image compression aims to improve compression efficiency by leveraging correlations between images. However, the requirement of synchronization and inter-image communication at the encoder side poses significant challenges, especially for constrained devices. In this study, we introduce a novel distributed image compression model based on the attention mechanism to address the challenges associated with the availability of side information only during decoding. Our model integrates an encoder network, a quantization module, and a decoder network, to ensure both high compression performance and high-quality image reconstruction. The encoder uses a deep Convolutional Neural Network (CNN) to extract high-level features from the input image, which then pass through the quantization module for further compression before undergoing lossless entropy coding. The decoder of our model consists of three main components that allow us to fully exploit the information within and between images on the decoder side. Specifically, we first introduce a channel-spatial attention module to capture and refine information within individual image feature maps. Second, we employ a semi-coupled convolution module to extract both shared and specific information in images. Finally, a cross-attention module is employed to fuse mutual information extracted from side information. The effectiveness of our model is validated on various datasets, including KITTI Stereo and Cityscapes. The results highlight the superior compression capabilities of our method, surpassing state-of-the-art techniques.http://www.sciencedirect.com/science/article/pii/S2352864824000683Digital communicationImage compressionSide informationChannel-spatial attention moduleCross-attention module
spellingShingle Yankai Yin
Zhe Sun
Peiying Ruan
Ruidong Li
Feng Duan
Learned distributed image compression with decoder side information
Digital Communications and Networks
Digital communication
Image compression
Side information
Channel-spatial attention module
Cross-attention module
title Learned distributed image compression with decoder side information
title_full Learned distributed image compression with decoder side information
title_fullStr Learned distributed image compression with decoder side information
title_full_unstemmed Learned distributed image compression with decoder side information
title_short Learned distributed image compression with decoder side information
title_sort learned distributed image compression with decoder side information
topic Digital communication
Image compression
Side information
Channel-spatial attention module
Cross-attention module
url http://www.sciencedirect.com/science/article/pii/S2352864824000683
work_keys_str_mv AT yankaiyin learneddistributedimagecompressionwithdecodersideinformation
AT zhesun learneddistributedimagecompressionwithdecodersideinformation
AT peiyingruan learneddistributedimagecompressionwithdecodersideinformation
AT ruidongli learneddistributedimagecompressionwithdecodersideinformation
AT fengduan learneddistributedimagecompressionwithdecodersideinformation