无需训练即可实现精准图像编辑，提升提示对齐与细节保真度。
FlowBypass: Rectified Flow Trajectory Bypass for Training-Free Image Editing
- 基于修正流构建绕行路径，避免误差累积
- 在多个数据集上优于当前最优方法，保持高保真度
- 适用于任意模型，无需特征调优，通用性强
训练无关的图像编辑因其高效性和不依赖训练数据而受到关注。然而，现有方法主要依赖反演-重建轨迹，存在固有权衡：轨迹越长，误差累积越多，影响保真度；越短则难以充分对齐编辑提示。此前方法多采用特定主干网络的特征操作，泛化性受限。为此，我们提出 FlowBypass，一种基于修正流的分析框架，直接构建反演与重建轨迹之间的绕行路径，从而在不依赖特征操作的前提下缓解误差累积。我们推导了两条轨迹的数学表达，获得近似绕行公式及其数值解，实现轨迹平滑过渡。大量实验表明，FlowBypass 持续优于当前最优图像编辑方法，在保证无关区域高保真细节的同时，实现更强的提示对齐能力。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。