BriefGPT.xyz
Oct, 2022
基于扩散的语义图像编辑与掩模引导
DiffEdit: Diffusion-based semantic image editing with mask guidance
HTML
PDF
Guillaume Couairon, Jakob Verbeek, Holger Schwenk, Matthieu Cord
TL;DR
本文提出了一种基于条件扩散模型的语义图像编辑方法DiffEdit,能够自动生成需要编辑的图像区域的遮罩,并利用潜在推理保留感兴趣区域的内容,该方法在ImageNet数据集上实现了最先进的编辑表现。
Abstract
Image generation has recently seen tremendous advances, with
diffusion models
allowing to synthesize convincing images for a large variety of
text prompts
. In this article, we propose DiffEdit, a method to take a
→