Conditional Video Generation for High-Efficiency Video Compression
This work addresses the problem of efficient video compression for applications requiring high perceptual quality, representing an incremental advance over existing neural codecs.
The paper tackles video compression by reframing it as a conditional generation task using diffusion models, achieving significant improvements in perceptual quality metrics like FVD and LPIPS, particularly at high compression ratios.
Perceptual studies demonstrate that conditional diffusion models excel at reconstructing video content aligned with human visual perception. Building on this insight, we propose a video compression framework that leverages conditional diffusion models for perceptually optimized reconstruction. Specifically, we reframe video compression as a conditional generation task, where a generative model synthesizes video from sparse, yet informative signals. Our approach introduces three key modules: (1) Multi-granular conditioning that captures both static scene structure and dynamic spatio-temporal cues; (2) Compact representations designed for efficient transmission without sacrificing semantic richness; (3) Multi-condition training with modality dropout and role-aware embeddings, which prevent over-reliance on any single modality and enhance robustness. Extensive experiments show that our method significantly outperforms both traditional and neural codecs on perceptual quality metrics such as Fréchet Video Distance (FVD) and LPIPS, especially under high compression ratios.