WAFI-VSC: Wireless Adaptive Frame Interpolation Video Semantic Communication

Semantic communication has emerged as one of the most promising paradigms for the upcoming 6G communication. In 6G communication, ultra-low latency video communication is necessary to meet the requirements of scenarios such as autonomous driving and virtual reality (VR). Existing video semantic communication schemes primarily focus on ensuring high video reconstruction quality, while the issue of latency remains largely overlooked. This paper proposes an end-to-end video semantic communication system for ultra-low latency communication, named wireless adaptive frame interpolation video semantic communication (WAFI-VSC). WAFI-VSC employs a simple key frame selector to efficiently choose the key frames that need to be encoded and transmitted. By utilizing deep learning-based frame interpolation technology, it achieves rapid video reconstruction. Furthermore, to ensure the reliability of the system, WAFI-VSC incorporates an adaptive group of picture (GOP) length mechanism based on channel conditions and video conditions. Experimental results show that compared to other semantic video communication and traditional video communication schemes, WAFI-VSC not only enhances video reconstruction quality but also markedly reduces processing latency and provides an efficient solution for ultra-low latency video semantic communication.