Enhancing Image Aesthetics with Dual-Conditioned Diffusion Models Guided by Multimodal Perception

ArXi:2603.11556v1 Announce Type: new Image aesthetic enhancement aims to perceive aesthetic deficiencies in images and perform corresponding editing operations, which is highly challenging and requires the model to possess creativity and aesthetic perception capabilities. Although recent advancements in image editing models have significantly enhanced their controllability and flexibility, they struggle with enhancing image aesthetic.