VQ-DeepVSC: A Dual-Stage Vector Quantization Framework for Video Semantic Communication

TL;DR

VQ-DeepVSC introduces a dual-stage vector quantization framework for video semantic communication, improving transmission quality over wireless channels by adaptive keyframe selection and efficient semantic compression, outperforming traditional standards especially in challenging conditions.

Contribution

The paper presents a novel dual-stage vector quantization framework with adaptive keyframe extraction and semantic encoding, enhancing wireless video transmission efficiency and robustness compared to existing methods.

Findings

Achieves higher MS-SSIM and LPIPS scores than H.265.

Demonstrates robustness under low SNR and multi-path channels.

Outperforms traditional joint source-channel coding frameworks.

Abstract

In response to the rapid growth of global videomtraffic and the limitations of traditional wireless transmission systems, we propose a novel dual-stage vector quantization framework, VQ-DeepVSC, tailored to enhance video transmission over wireless channels. In the first stage, we design the adaptive keyframe extractor and interpolator, deployed respectively at the transmitter and receiver, which intelligently select key frames to minimize inter-frame redundancy and mitigate the cliff-effect under challenging channel conditions. In the second stage, we propose the semantic vector quantization encoder and decoder, placed respectively at the transmitter and receiver, which efficiently compress key frames using advanced indexing and spatial normalization modules to reduce redundancy. Additionally, we propose adjustable index selection and recovery modules, enhancing compression efficiency and enabling flexible compression ratio adjustment. Compared to the joint source-channel coding (JSCC) framework, the proposed framework exhibits superior compatibility with current digital communication systems. Experimental results demonstrate that VQ-DeepVSC achieves substantial improvements in both Multi-Scale Structural Similarity (MS-SSIM) and Learned Perceptual Image Patch Similarity (LPIPS) metrics than the H.265 standard, particularly under low channel signal-to-noise ratio (SNR) or multi-path channels, highlighting the significantly enhanced transmission capabilities of our approach.