Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-hop Parallel Image Semantic Communication for Distortion Accumulation Mitigation

Bingyan Xie, Jihong Park, Yongpeng Wu, Wenjun Zhang, Tony Quek

TL;DR

The paper tackles distortion accumulation in multi-hop wireless image transmission for semantic communication by introducing Multi-hop Parallel Semantic Communication (MHPSC) with a parallel residual compensation link. It combines a coarse-to-fine residual compression pipeline—consisting of a DL residual compressor and adaptive arithmetic coding (AAC) guided by a learned residual distribution estimator—to transmit residuals with minimal bandwidth overhead. Key contributions include the residual estimation module modeling the residual distribution as a mixture of logistic components, a two-stream MHPSC architecture with a compensation path, and a staged training paradigm that enhances multi-hop robustness. Experimental results on the UDIS-D dataset show MHPSC outperforms existing semantic and traditional SSCC schemes under various SNRs, CBRs, and hop counts, with only a modest bandwidth increase, indicating strong practical potential for reliable multi-hop image transmission in 6G-type networks.

Abstract

Existing semantic communication schemes primarily focus on single-hop scenarios, overlooking the challenges of multi-hop wireless image transmission. As semantic communication is inherently lossy, distortion accumulates over multiple hops, leading to significant performance degradation. To address this, we propose the multi-hop parallel image semantic communication (MHPSC) framework, which introduces a parallel residual compensation link at each hop against distortion accumulation. To minimize the associated transmission bandwidth overhead, a coarse-to-fine residual compression scheme is designed. A deep learning-based residual compressor first condenses the residuals, followed by the adaptive arithmetic coding (AAC) for further compression. A residual distribution estimation module predicts the prior distribution for the AAC to achieve fine compression performances. This approach ensures robust multi-hop image transmission with only a minor increase in transmission bandwidth. Experimental results confirm that MHPSC outperforms both existing semantic communication and traditional separated coding schemes.

Multi-hop Parallel Image Semantic Communication for Distortion Accumulation Mitigation

TL;DR

The paper tackles distortion accumulation in multi-hop wireless image transmission for semantic communication by introducing Multi-hop Parallel Semantic Communication (MHPSC) with a parallel residual compensation link. It combines a coarse-to-fine residual compression pipeline—consisting of a DL residual compressor and adaptive arithmetic coding (AAC) guided by a learned residual distribution estimator—to transmit residuals with minimal bandwidth overhead. Key contributions include the residual estimation module modeling the residual distribution as a mixture of logistic components, a two-stream MHPSC architecture with a compensation path, and a staged training paradigm that enhances multi-hop robustness. Experimental results on the UDIS-D dataset show MHPSC outperforms existing semantic and traditional SSCC schemes under various SNRs, CBRs, and hop counts, with only a modest bandwidth increase, indicating strong practical potential for reliable multi-hop image transmission in 6G-type networks.

Abstract

Existing semantic communication schemes primarily focus on single-hop scenarios, overlooking the challenges of multi-hop wireless image transmission. As semantic communication is inherently lossy, distortion accumulates over multiple hops, leading to significant performance degradation. To address this, we propose the multi-hop parallel image semantic communication (MHPSC) framework, which introduces a parallel residual compensation link at each hop against distortion accumulation. To minimize the associated transmission bandwidth overhead, a coarse-to-fine residual compression scheme is designed. A deep learning-based residual compressor first condenses the residuals, followed by the adaptive arithmetic coding (AAC) for further compression. A residual distribution estimation module predicts the prior distribution for the AAC to achieve fine compression performances. This approach ensures robust multi-hop image transmission with only a minor increase in transmission bandwidth. Experimental results confirm that MHPSC outperforms both existing semantic communication and traditional separated coding schemes.

Paper Structure

This paper contains 18 sections, 14 equations, 5 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Proposed Framework
  3. Common Multi-hop Wireless Image Transmission Model
  4. Proposed Multi-hop Parallel Image Semantic Communication Framework
  5. Residual Estimation
  6. Deployment Detail
  7. Network Structure
  8. Training Loss and Strategy
  9. Numerical Results
  10. Experimental Setups
  11. Datasets
  12. Model Deployment Details
  13. Comparison Benchmarks
  14. Evaluation Metrics
  15. Results Analysis
  16. ...and 3 more sections

Figures (5)

  • Figure 1: The system model for the multi-hop wireless image transmission scenario. (a) The common multi-hop wireless image transmission link. (b) The proposed MHPSC with parallel link for multi-hop compensation. The red lines are the residual transmission process.
  • Figure 2: The residual estimation module for predicting distributions.
  • Figure 3: Quality of the reconstructed images versus the SNRs under Rayleigh fading channels (CBR = 0.071, N = 20).
  • Figure 4: Quality of the reconstructed images versus the CBRs under Rayleigh fading channels (SNR = 10 dB, N = 20).
  • Figure 5: Performance of different hop numbers.