面向大规模分层图像生成与编辑的掩码区域 Transformer
Masked Region Transformer for Layered Image Generation and Editing at Scale
把「选择性词元掩码」作为统一接口，让 text-to-layers、image-to-layers、layers-to-layers 共享同一个扩散框架，可直接迁移到需要分层/可编辑输出的生成任务；overflow-aware canvas layer 处理超出画布边界的图层与半透明背景合成，是分层生成中值得复用的表示技巧。
原文依据 we unify three complementary tasks--text-to-layers, image-to-layers, and layers-to-layers--within a shared masked region diffusion framework
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论分层图像生成与编辑是一项基础能力，它使生成的视觉内容能够像自然语言中的词级编辑那样，实现逐层的复用、编辑与组合。尽管其重要性不言而喻，这一方向在大规模层面仍鲜有探索。为填补这一空白，我们提出 MRT，一个 200 亿参数的掩码区域扩散模型，专为多层透明图像生成与编辑而设计，并在超过 1000 万个多语言设计样本上训练，覆盖多种宽高比与文本提示。为充分利用这一规模，我们做出两项关键技术贡献。第一，我们在一个共享的掩码区域扩散框架内统一了三个互补任务——文本到图层、图像到图层、图层到图层，其中选择性词元掩码使灵活的逐层生成与编辑成为可能。第二，为实现溢出图层生成，我们引入溢出感知画布图层，用于处理边界不一致问题并支持半透明背景合成，从而生成可超出可见画布边界的完整可编辑图层。此外，我们采用扩散蒸馏，以极小的质量损失实现 8 步实时多层生成。大量实验表明，我们的框架在全部三个任务上均显著优于此前的最优方法（包括多种商业系统），为多层透明图像生成建立了新的基准。值得注意的是，根据用户研究结果，我们的模型在图像到图层质量上显著优于同期提出的 Qwen-Image-Layered 模型，同时推理速度快 10~100 倍，并在图像到图层推理过程中节省 50%~90% 的激活显存占用。
英文原文摘要 1736 字符
Layered image generation and editing is a fundamental capability that enables layer-wise reuse, editing, and composition of generated visual content, analogous to word-level editing in natural language. Despite its importance, this remains an underexplored area at scale. To address this gap, we present MRT, a 20B-parameter masked region diffusion model tailored for multi-layer transparent image generation and editing, trained on over 10M multilingual design samples spanning diverse aspect ratios and textual prompts. To fully leverage this scale, we make two key technical contributions. First, we unify three complementary tasks--text-to-layers, image-to-layers, and layers-to-layers--within a shared masked region diffusion framework, where selective token masking enables flexible layer-wise generation and editing. Second, to enable overflow layer generation, we introduce an overflow-aware canvas layer that handles boundary inconsistencies and supports semi-transparent background synthesis, enabling complete editable layers extending beyond visible canvas boundaries. Additionally, we apply diffusion distillation to achieve 8-step, real-time multi-layer generation with minimal quality degradation. Extensive experiments demonstrate that our framework substantially outperforms prior state-of-the-art approaches, including various commercial systems, across all three tasks, establishing a new benchmark for multi-layer transparent image generation. Notably, our model significantly outperforms the concurrent Qwen-Image-Layered model in image-to-layers quality according to user-study results, while achieving 10~100xfaster inference and saving a 50%~90% activation GPU memory consumption during image-to-layer inference.