arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2507.01496cs.CV

ReFlex:通过中间步骤特征提取与注意力适配实现校正流中真实图像的文本引导编辑

ReFlex: Text-Guided Editing of Real Images in Rectified Flow via Mid-Step Feature Extraction and Attention Adaptation

Department of Intelligence and Information(智能与信息系)

Jimyeong Kim, Jungwon Park, Yeji Song, Nojun Kwak, Wonjong Rhee

AI总结:

本文提出ReFlex方法,通过分析多模态Transformer中间表示并利用中间步骤潜变量与注意力适配,实现无需训练和掩码的校正流真实图像文本引导编辑,实验证明其优于现有方法。

摘要:

校正流文本到图像模型在图像质量和文本对齐方面超越了扩散模型,但将ReFlow适配于真实图像编辑仍然具有挑战性。我们通过分析多模态Transformer块的中间表示并识别出三个关键特征,提出了一种新的面向ReFlow的真实图像编辑方法。为了在充分保留结构的前提下从真实图像中提取这些特征,我们利用了仅反演至中间步骤的中间步骤潜变量。随后,我们在注入过程中适配注意力,以提高可编辑性并增强与目标文本的对齐。我们的方法无需训练,不需要用户提供掩码,甚至在没有源提示的情况下也能应用。在两个基准上与九种基线方法进行的大量实验表明,其性能优于先前方法,人类评估进一步验证了用户对我们方法的强烈偏好。

英文摘要:

Rectified Flow text-to-image models surpass diffusion models in image quality and text alignment, but adapting ReFlow for real-image editing remains challenging. We propose a new real-image editing method for ReFlow by analyzing the intermediate representations of multimodal transformer blocks and identifying three key features. To extract these features from real images with sufficient structural preservation, we leverage mid-step latent, which is inverted only up to the mid-step. We then adapt attention during injection to improve editability and enhance alignment to the target text. Our method is training-free, requires no user-provided mask, and can be applied even without a source prompt. Extensive experiments on two benchmarks with nine baselines demonstrate its superior performance over prior methods, further validated by human evaluations confirming a strong user preference for our approach.

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。

↑