Lightweight Vision Model-based Multi-user Semantic Communication Systems

TL;DR

The paper tackles bandwidth-constrained image SemCom by introducing LVM-MSC, a lightweight multi-user SemCom framework. It integrates a Fast SAM-based Lightweight Knowledge Base (LKB) to rapidly locate semantic objects, an MAE-based Efficient Semantic Codec (ESC) that adaptively masks background pixels to maximize semantic density, and a Multi-user Semantic Sharing (MSS) scheme that exploits shared semantics across users to broadcast common content. Training proceeds in three phases with specialized losses, while an asymmetric encoder/decoder design places the heavy computation at the base station and lightweight decoding at user devices. Simulation results on PASCAL VOC2012 and CIFAR-10 demonstrate reduced bandwidth, improved region-specific reconstruction (PSNR/SSIM), and superior classification performance under varying channel conditions, highlighting practical gains for edge-enabled SemCom. Overall, LVM-MSC offers a scalable approach to efficient, multi-user SemCom by combining lightweight knowledge representation, targeted semantic encoding, and semantic-space sharing.

Abstract

Semantic Communication (SemCom) is a promising new paradigm for next-generation communication systems, emphasizing the transmission of core information, particularly in environments characterized by uncertainty, noise, and bandwidth constraints. However, existing image SemCom systems face several challenges, such as inefficient knowledge base construction, insufficient semantic encoding, and lack of multi-user semantic sharing. To address these issues, we propose a Lightweight Vision Model-based Multi-user Semantic Communication System (LVM-MSC). First, we construct a Lightweight Knowledge Base (LKB) based on the fast Segment Anything Model (SAM). LKB incorporates the extensive image knowledge of the SAM model while significantly reducing the number of parameters through its convolutional architecture. Next, we design an Efficient Semantic Codec (ESC) based on the Masked AutoEncoder (MAE) architecture. ESC enhances semantic compression at both the pixel and semantic levels and implements lightweight semantic decoding tailored for user devices. Furthermore, we propose a Multi-user Semantic Sharing (MSS) transmission for the multi-user SemCom. By calculating the similarity of semantic information among different users in the sharing semantic space, we unify the transmissions of similar semantic information through broadcasting, further improving the transmission efficiency. Finally, simulation results demonstrate the feasibility and effectiveness of the proposed LVM-MSC system.

Figures (11)

• Figure 1: The structure of system model.
• Figure 2: The structure of the proposed SemCom system.
• Figure 3: Training of the proposed SemCom system.
• Figure 4: The workflow of LKB.
• Figure 5: The workflow of ESC.