DiTVR: Zero-Shot Diffusion Transformer for Video Restoration
This addresses the problem of generating high-quality videos from low-quality inputs for applications like super-resolution and denoising, though it is incremental by building on existing diffusion and transformer methods.
The paper tackled video restoration by introducing DiTVR, a zero-shot diffusion transformer framework that achieved new state-of-the-art results on benchmarks, with improved temporal consistency and detail preservation.
Video restoration aims to reconstruct high quality video sequences from low quality inputs, addressing tasks such as super resolution, denoising, and deblurring. Traditional regression based methods often produce unrealistic details and require extensive paired datasets, while recent generative diffusion models face challenges in ensuring temporal consistency. We introduce DiTVR, a zero shot video restoration framework that couples a diffusion transformer with trajectory aware attention and a wavelet guided, flow consistent sampler. Unlike prior 3D convolutional or frame wise diffusion approaches, our attention mechanism aligns tokens along optical flow trajectories, with particular emphasis on vital layers that exhibit the highest sensitivity to temporal dynamics. A spatiotemporal neighbour cache dynamically selects relevant tokens based on motion correspondences across frames. The flow guided sampler injects data consistency only into low-frequency bands, preserving high frequency priors while accelerating convergence. DiTVR establishes a new zero shot state of the art on video restoration benchmarks, demonstrating superior temporal consistency and detail preservation while remaining robust to flow noise and occlusions.