arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2412.21079cs.CV
收藏

Edicho:自然场景下的一致性图像编辑

Edicho: Consistent Image Editing in the Wild

HKUST(香港科技大学);Ant Group(蚂蚁集团);Stanford University(斯坦福大学);CUHK(香港中文大学)

Qingyan Bai, Hao Ouyang, Yinghao Xu, Qiuyu Wang, Ceyuan Yang, Ka Leong Cheng, Yujun Shen, Qifeng Chen

AI总结:

针对自然场景跨图一致性编辑的技术难题,提出免训练的Edicho方法,基于显式图像对应关系设计注意力操控模块与优化的CFG去噪策略,即插即用且兼容主流扩散编辑方法,实验验证了其有效性。

摘要:

经证实存在实际需求的自然场景图像跨图一致性编辑,因物体姿态、光照条件、拍摄环境等多种不可控因素,仍是一项技术挑战。Edicho提出了一种基于扩散模型的免训练解决方案,其核心设计原则是利用显式图像对应关系引导编辑。具体而言,关键组件包括注意力操控模块和经过精心优化的无分类器引导(CFG)去噪策略,二者均结合了预先估计的对应关系。这种推理阶段算法具备即插即用特性,可兼容大多数基于扩散的编辑方法,如ControlNet和BrushNet。大量实验结果证明了Edicho在多种场景下跨图一致性编辑的有效性。我们将发布代码以推动后续研究。

英文摘要:

As a verified need, consistent editing across in-the-wild images remains a technical challenge arising from various unmanageable factors, like object poses, lighting conditions, and photography environments. Edicho steps in with a training-free solution based on diffusion models, featuring a fundamental design principle of using explicit image correspondence to direct editing. Specifically, the key components include an attention manipulation module and a carefully refined classifier-free guidance (CFG) denoising strategy, both of which take into account the pre-estimated correspondence. Such an inference-time algorithm enjoys a plug-and-play nature and is compatible to most diffusion-based editing methods, such as ControlNet and BrushNet. Extensive results demonstrate the efficacy of Edicho in consistent cross-image editing under diverse settings. We will release the code to facilitate future studies.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑