用掩码生成 Transformer 实现精准图像编辑，避免修改扩散到无关区域。
Masked Generative Transformer Is What You Need for Image Editing
- 采用局部令牌预测机制，限制修改仅作用于目标区域。
- 在多个基准上达到领先图像相似度，速度比扩散模型快6倍。
- 适合追求高效精准编辑的开发者与研究者使用。
扩散模型主导图像编辑，但其全局去噪机制会将修改传播至不应变动的邻近区域。本文提出基于掩码生成Transformer（MGT）的全新方法，其局部令牌预测范式天然限制修改范围。我们构建了首个基于MGT的编辑框架EditMGT，通过多层注意力整合，将跨注意力图聚合为精确的编辑定位信号，并采用区域保持采样，显式防止非目标区域的令牌翻转。为支持训练，我们创建了包含200万样本、分辨率超1024的CrispEdit-2M数据集，覆盖七个类别。仅960M参数的EditMGT在多个基准上实现最优图像相似度，且编辑速度提升6倍，证明MGT是扩散模型的有力替代方案。
原文摘要 · Abstract (English)
Diffusion models dominate image editing, yet their global denoising mechanism entangles edited regions with surrounding context, causing modifications to propagate into areas that should remain intact. We propose a fundamentally different approach by leveraging Masked Generative Transformers (MGTs), whose localized token-prediction paradigm naturally confines changes to intended regions. We present EditMGT, an MGT-based editing framework that is the first of its kind. Our approach employs multi-layer attention consolidation to aggregate cross-attention maps into precise edit localization signals, and region-hold sampling to explicitly prevent token flipping in non-target areas. To support training, we construct CrispEdit-2M, a 2M-sample high-resolution (>1024) editing dataset spanning seven categories. With only 960M parameters, EditMGT achieves state-of-the-art image similarity on multiple benchmarks while delivering 6x faster editing, demonstrating that MGTs offer a compelling alternative to diffusion-based editing.
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。