FlowBypass: 基于修正流的无训练图像编辑轨迹绕过
FlowBypass: Rectified Flow Trajectory Bypass for Training-Free Image Editing
AI总结：
FlowBypass通过基于修正流的轨迹绕过方法，在无需特征操作的情况下提升无训练图像编辑的保真度与提示对齐能力。
摘要：
无训练图像编辑因其高效性和对训练数据的独立性而受到越来越多的关注。然而，现有方法大多依赖于逆向-重建轨迹，这带来了固有的权衡：更长的轨迹会积累误差并影响保真度，而更短的轨迹则无法确保充分与编辑提示对齐。以往尝试解决这一问题的方法通常采用特定骨干网络的特征操作，限制了通用性。为了解决这些挑战，我们提出了FlowBypass，一种新颖且分析性的框架，基于修正流，构建直接连接逆向和重建轨迹的绕过路径，从而在不依赖特征操作的情况下缓解误差积累。我们提供了两个轨迹的正式推导，从中获得了近似绕过公式及其数值解，使轨迹转换得以无缝进行。大量实验表明，FlowBypass在性能上始终优于最先进图像编辑方法，在保持无关区域高保真细节的同时，实现了更强的提示对齐。
英文摘要：
Training-free image editing has attracted increasing attention for its efficiency and independence from training data. However, existing approaches predominantly rely on inversion-reconstruction trajectories, which impose an inherent trade-off: longer trajectories accumulate errors and compromise fidelity, while shorter ones fail to ensure sufficient alignment with the edit prompt. Previous attempts to address this issue typically employ backbone-specific feature manipulations, limiting general applicability. To address these challenges, we propose FlowBypass, a novel and analytical framework grounded in Rectified Flow that constructs a bypass directly connecting inversion and reconstruction trajectories, thereby mitigating error accumulation without relying on feature manipulations. We provide a formal derivation of two trajectories, from which we obtain an approximate bypass formulation and its numerical solution, enabling seamless trajectory transitions. Extensive experiments demonstrate that FlowBypass consistently outperforms state-of-the-art image editing methods, achieving stronger prompt alignment while preserving high-fidelity details in irrelevant regions.
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。
