Distributed Image Semantic Communication via Nonlinear Transform Coding

TL;DR

This paper introduces a novel distributed image semantic communication framework using nonlinear transform coding, explicitly modeling source correlation to improve transmission efficiency and quality over wireless channels.

Contribution

It proposes a new approach leveraging nonlinear transform coding for distributed image semantic communication, explicitly modeling source correlation, and implementing it with Swin Transformers for enhanced performance.

Findings

Outperforms existing distributed SSCC and JSCC methods.

Achieves state-of-the-art results in pixel-level and perceptual quality.

Effectively models source correlation for improved transmission.

Abstract

This paper investigates distributed source-channel coding for correlated image semantic transmission over wireless channels. In this setup, correlated images at different transmitters are separately encoded and transmitted through dedicated channels for joint recovery at the receiver. We propose a general approach for distributed image semantic communication that applies to both separate source and channel coding (SSCC) and joint source-channel coding (JSCC). Unlike existing learning-based approaches that implicitly learn source correlation in a purely data-driven manner, our method leverages nonlinear transform coding (NTC) to explicitly model source correlation from both probabilistic and geometric perspectives. A joint entropy model approximates the joint distribution of latent representations to guide adaptive rate allocation, while a transformation module aligns latent features for maximal correlation learning at the decoder. We implement this framework as D-NTSC for SSCC and D-NTSCC for JSCC, both built on Swin Transformers for effective feature extraction and correlation exploitation. Variational inference is employed to derive principled loss functions that jointly optimize encoding, decoding, and joint entropy modeling. Extensive experiments on real-world multi-view datasets demonstrate that D-NTSC and D-NTSCC outperform existing distributed SSCC and distributed JSCC baselines, respectively, achieving state-of-the-art performance in both pixel-level and perceptual quality metrics.