Joint Semantic Feature Selection and Bandwidth Allocation for Image Transmission Optimization
Chen Chen, Yang Li, Li Feng · 2024
Semantic communication, which exploits advanced deep neural networks to achieve semantic understanding, has developed as a novel approach to conveying semantic information rather than the source data. Different from traditional communication methods, semantic communication focuses on the information content and extracts key semantic features for transmission. Specifically, a smaller bandwidth resource with a higher transmission efficiency can be achieved in semantic communication. With these benefits, semantic communication has been considered a promising solution, which enables ultralow latency and high-quality image transmission. Most current works focus on the design of semantic encoder and decoder while neglecting the transmission efficiency of semantic features. Thus, in this paper, we consider the impact of different features on the performance of latency and image quality. To this end, a joint optimization of semantic feature selection and bandwidth allocation is formulated to minimize the overall latency with the constraint of image quality loss. Moreover, the Harris Hawk optimization algorithm is adopted to find the optimal feature selections. Then, with the given selected features, it can be proved that the overall latency minimization is a convex optimization problem, and the Karush-Kuhn-Tucker condition is used to calculate the optimal bandwidth allocation. Numerical experiment results are provided to demonstrate the effectiveness and efficiency of the proposed image semantic transmission scheme.