整流流图像编辑的最优传输：统一基于反演与直接方法
Optimal Transport for Rectified Flow Image Editing: Unifying Inversion-Based and Direct Methods
AI总结：
本文提出统一最优传输框架，同时改进整流流图像编辑中基于反演和无需求逆两类方法，在重建保真度、结构保持与编辑可控性上均取得显著提升。
摘要：
整流流模型中的图像编辑仍然具有挑战性，原因在于重建保真度与编辑灵活性之间的根本性权衡。基于反演的方法受到轨迹偏差的影响，而近期无需求逆的方法（如FlowEdit）虽然提供了直接的编辑路径，但可以通过额外的引导来改善结构保持。在本工作中，我们证明最优传输理论为改进整流流编辑中的这两种范式提供了一个统一框架。我们引入了一个零样本的传输引导反演框架，该框架在反向扩散过程中利用最优传输，并将最优传输原理扩展到通过传输优化的速度场校正来增强无需求逆的方法。引入基于传输的引导能够在不同的整流流编辑方法中有效平衡重建精度与编辑可控性。对于基于反演的编辑，我们的方法在人脸编辑基准上实现了高保真重建，LPIPS分数为0.001，SSIM为0.992，在LSUN数据集上观察到相比RF-Inversion有7.8%到12.9%的提升。对于在FLUX和Stable Diffusion 3上使用FlowEdit的无需求逆编辑，我们在多种编辑场景中展示了一致的语义一致性和结构保持改进。我们的语义人脸编辑实验显示身份保持提升了11.2%，感知质量也有所增强。该统一最优传输框架在基于反演和直接编辑两种范式中都产生了视觉上引人注目的编辑结果，并具有优越的细节保持能力。代码可在RF-Inversion和FlowEdit获取：https://github.com/marianlupascu/OT-RF
英文摘要：
Image editing in rectified flow models remains challenging due to the fundamental trade-off between reconstruction fidelity and editing flexibility. While inversion-based methods suffer from trajectory deviation, recent inversion-free approaches like FlowEdit offer direct editing pathways but can benefit from additional guidance to improve structure preservation. In this work, we demonstrate that optimal transport theory provides a unified framework for improving both paradigms in rectified flow editing. We introduce a zero-shot transport-guided inversion framework that leverages optimal transport during the reverse diffusion process, and extend optimal transport principles to enhance inversion-free methods through transport-optimized velocity field corrections. Incorporating transport-based guidance can effectively balance reconstruction accuracy and editing controllability across different rectified flow editing approaches. For inversion-based editing, our method achieves high-fidelity reconstruction with LPIPS scores of 0.001 and SSIM of 0.992 on face editing benchmarks, observing 7.8% to 12.9% improvements over RF-Inversion on LSUN datasets. For inversion-free editing with FlowEdit on FLUX and Stable Diffusion 3, we demonstrate consistent improvements in semantic consistency and structure preservation across diverse editing scenarios. Our semantic face editing experiments show an 11.2% improvement in identity preservation and enhanced perceptual quality. The unified optimal transport framework produces visually compelling edits with superior detail preservation across both inversion-based and direct editing paradigms. Code is available for RF-Inversion and FlowEdit at: https://github.com/marianlupascu/OT-RF
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。
