A Wyner-Ziv Coding-Based Semantic Communication Approach with a Shared Semantic Codebook.

Recently, a codebook-assisted semantic communication framework has demonstrated performance gain over traditional semantic communication systems. In this paper, we further improve the performance of this method for image semantic communication from two aspects. First, we propose a semantic-aware codebook construction method, which is based on a novel distance measure, unifying the targets of the codeword encoder and the NN - based semantic encoder. Second, to enhance the transmission efficiency, we design the NN-based Wyner-Ziv encoder, which aims to reduce the redundancy between the coding results of the NN-based semantic encoder and the corresponding codewords. Numerical results show that our method can achieve a better performance than the conventional and codebook-assisted semantic encoding methods in both semantic information recovery and data information recovery.