Video Coding

Video coding aims to efficiently compress and represent video data for storage and transmission, balancing compression ratio with reconstruction quality for both human viewing and machine analysis. Current research emphasizes developing faster encoding/decoding methods, often leveraging neural networks (e.g., transformers, GANs, diffusion models) and incorporating coding priors (motion vectors, residual frames) to improve efficiency and perceptual quality. These advancements are crucial for managing the ever-increasing volume of video data and enabling real-time applications like live video analytics and machine vision tasks, impacting both scientific understanding of compression and practical deployment in various industries.

Papers