Decoupled Data Consistency with Diffusion Purification for Image Restoration

arXiv:2403.06054v1 Announce Type: cross
Abstract: Diffusion models have recently gained traction as a powerful class of deep generative priors, excelling in a wide range of image restoration tasks due to their exceptional ability to model data distributions. To solve image restoration problems, many existing techniques achieve data consistency by incorporating additional likelihood gradient steps into the reverse sampling process of diffusion models. However, the additional gradient steps pose a challenge for real-world practical applications as they incur a large computational …

abstract arxiv class cs.ai cs.cv cs.lg data diffusion diffusion models eess.iv eess.sp generative image image restoration solve tasks type