-8 dB SNR + 90% Packet Loss: MamVSC -- CSI-Guided Semantic Mamba for Extreme-Robust Video Semantic Communication

2026-07-08

AI Summary

This work proposes MamVSC, a novel semantic communication system for wireless video transmission that addresses two critical distortion types: semantic shift caused by channel noise and semantic erasure due to high packet loss. MamVSC is the first to jointly model these distortions, introducing an innovative CSI-guided Mamba-based semantic extraction module and a dynamic semantic clustering center coding-decoding mechanism. These components enable adaptive adjustment of semantic granularity and optimized recovery strategies tailored to channel conditions. Even under extremely harsh channel environments—specifically, AWGN with −8 dB SNR and 90% packet loss—the system achieves high-quality video reconstruction with MS-SSIM exceeding 0.6 and PSNR above 21 dB, demonstrating significantly enhanced robustness in semantic communication.

Abstract

Semantic communication, leveraging joint source-channel coding, is designed to mitigate semantic distortion introduced by the channel. However, most current studies focus solely on semantic deviation distortion caused by physical wireless channels, while overlooking semantic erasure distortion due to packet loss. A CSI-Guided Mamba-based video semantic wireless digital communication system (MamVSC) employing semantic grouping is proposed to simultaneously address both semantic deviation and erasure distortions. In this system, a semantic Mamba module, guided by channel state information (CSI) feedback, is utilized to dynamically adjust the granularity of extracted semantic information, adapting to channel conditions. Furthermore, a Semantic Channel Codec based on dynamic Semantic clustering centers is introduced, where the distance between semantic vectors within the same semantic class and their corresponding Semantic clustering center is dynamically adjusted according to channel conditions, enhancing robustness against channel noise. Additionally, a adaptive packet loss recovery module, dynamically adaptive to the CSI, is proposed. The system achieves an MS-SSIM greater than 0.6 and a PSNR exceeding 21 dB at an SNR of -8 dB and a packet loss rate of 90% in AWGN channel.

Problem

• semantic communication
• semantic distortion
• packet loss
• channel noise
• video transmission

Innovation

• Semantic Communication
• Mamba Architecture
• Channel State Information (CSI)
• Packet Loss Recovery
• Semantic Clustering