Diffusion Model Based Visual Compensation Guidance and Visual Difference Analysis for No-Reference Image Quality Assessment

arXiv:2402.14401v1 Announce Type: cross
Abstract: Existing free-energy guided No-Reference Image Quality Assessment (NR-IQA) methods still suffer from finding a balance between learning feature information at the pixel level of the image and capturing high-level feature information and the efficient utilization of the obtained high-level feature information remains a challenge. As a novel class of state-of-the-art (SOTA) generative model, the diffusion model exhibits the capability to model intricate relationships, enabling a comprehensive understanding of images and possessing a better learning of …

abstract analysis arxiv assessment balance compensation cs.cv cs.lg difference diffusion diffusion model eess.iv energy feature free guidance image information pixel quality reference type visual