标题： DNAEdit：文本引导的修正流编辑中的直接噪声对齐 显示英文标题

标题： DNAEdit: Direct Noise Alignment for Text-Guided Rectified Flow Editing

摘要： 利用大规模预训练文本到图像模型强大的生成能力，无训练方法已经展示了令人印象深刻的图像编辑结果。 基于扩散的传统方法以及最近基于修正流（RF）的方法通常通过逐渐向干净图像添加噪声来逆向合成轨迹，在此过程中，当前时间步的噪声潜在表示用于近似下一时间步的噪声潜在表示，从而引入累积漂移并降低重建准确性。 考虑到在RF中，噪声潜在表示是通过对每个时间步的高斯噪声和干净图像进行直接插值估计的这一事实，我们提出了直接噪声对齐（DNA），它直接在噪声域中精化所需的高斯噪声，显著减少了先前方法中的误差累积。 具体而言，DNA在每个时间步估计插值噪声潜在表示的速度场，并通过预测速度场与预期速度场之间的差异调整高斯噪声。 我们验证了DNA的有效性，并揭示了其与现有基于RF的反演方法的关系。 此外，我们引入了移动速度引导（MVG）以控制目标提示引导的生成过程，平衡图像背景保留和目标对象可编辑性。 DNA和MVG共同构成了我们提出的方法，即DNAEdit。 最后，我们介绍了DNA-Bench，一个长提示基准，用于评估先进图像编辑模型的性能。 实验结果表明，我们的DNAEdit在文本引导的编辑方法中取得了优于现有技术的性能。 代码和基准将在\href{ https://xiechenxi99.github.io/DNAEdit/}{https://xiechenxi99.github.io/DNAEdit/}处提供。 显示英文摘要

摘要： Leveraging the powerful generation capability of large-scale pretrained text-to-image models, training-free methods have demonstrated impressive image editing results. Conventional diffusion-based methods, as well as recent rectified flow (RF)-based methods, typically reverse synthesis trajectories by gradually adding noise to clean images, during which the noisy latent at the current timestep is used to approximate that at the next timesteps, introducing accumulated drift and degrading reconstruction accuracy. Considering the fact that in RF the noisy latent is estimated through direct interpolation between Gaussian noises and clean images at each timestep, we propose Direct Noise Alignment (DNA), which directly refines the desired Gaussian noise in the noise domain, significantly reducing the error accumulation in previous methods. Specifically, DNA estimates the velocity field of the interpolated noised latent at each timestep and adjusts the Gaussian noise by computing the difference between the predicted and expected velocity field. We validate the effectiveness of DNA and reveal its relationship with existing RF-based inversion methods. Additionally, we introduce a Mobile Velocity Guidance (MVG) to control the target prompt-guided generation process, balancing image background preservation and target object editability. DNA and MVG collectively constitute our proposed method, namely DNAEdit. Finally, we introduce DNA-Bench, a long-prompt benchmark, to evaluate the performance of advanced image editing models. Experimental results demonstrate that our DNAEdit achieves superior performance to state-of-the-art text-guided editing methods. Codes and benchmark will be available at \href{ https://xiechenxi99.github.io/DNAEdit/}{https://xiechenxi99.github.io/DNAEdit/}.