Semantics-Guided Contrastive Joint Source-Channel Coding for Image Transmission

Wenhui Hua, Dezhao Chen, Junli Fang, Lingyu Chen, João F. C. Mota, Xuemin Hong

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)
47 Downloads (Pure)

Abstract

Deepjoint source-channel coding (D-JSCC) provides several advantages over conventional coding schemes, in which source and channel coding are designed separately. For example, D-JSCC schemes suffer from smaller delays and are more robust to rapid channel variation. However, D-JSCCs are often designed without explicit structure or insight, making them less adaptive, hard to control, and theoretically unfounded. In this paper, we propose a contrastive joint-source-channel coding (C-JSCC) design, which uses supervised contrastive learning (SCL) to make the latent space of a conventional D-JSCC more structured and meaningful. By testing on the CIFAR-10 dataset, we show that C-JSCC consistently outperforms its D-JSCC counterpart in both tasks of image reconstruction and classification. Moreover, C-JSCC is shown to output images with perceptual quality better than the classic BPG image codec in the low bits-per-pixel (bpp) region. The roles of various hyper-parameters in C-JSCC are investigated by analytical approximations, experiments, and visualization techniques.
Original languageEnglish
Title of host publication14th International Conference on Wireless Communications and Signal Processing 2022
PublisherIEEE
Pages505-510
Number of pages6
ISBN (Electronic)9781665450850
DOIs
Publication statusPublished - 15 Feb 2023
Event14th International Conference on Wireless Communications and Signal Processing 2022 - Nanjing, China
Duration: 1 Nov 20223 Nov 2022

Conference

Conference14th International Conference on Wireless Communications and Signal Processing 2022
Abbreviated titleWCSP 2022
Country/TerritoryChina
CityNanjing
Period1/11/223/11/22

Keywords

  • Contrastive learning
  • image communications
  • joint source-channel coding
  • semantic information

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Signal Processing
  • Information Systems and Management
  • Safety, Risk, Reliability and Quality

Fingerprint

Dive into the research topics of 'Semantics-Guided Contrastive Joint Source-Channel Coding for Image Transmission'. Together they form a unique fingerprint.

Cite this