An Item is Worth a Prompt: Versatile Image Editing with Disentangled Control

arXiv:2403.04880v1 Announce Type: new
Abstract: Building on the success of text-to-image diffusion models (DPMs), image editing is an important application to enable human interaction with AI-generated content. Among various editing methods, editing within the prompt space gains more attention due to its capacity and simplicity of controlling semantics. However, since diffusion models are commonly pretrained on descriptive text captions, direct editing of words in text prompts usually leads to completely different generated images, violating the requirements for image editing. On …

abstract ai-generated content application arxiv attention building capacity control cs.cv diffusion diffusion models editing generated human image image diffusion prompt semantics simplicity space success text text-to-image the prompt type