UniWorld-Design：从像素生成到图层原生设计

摘要

我们提出 UniWorld-Design，一个将图像生成从平面像素合成重新定义为结构化视觉合成的框架，以语义 RGBA 图层作为生成、理解和编辑的原子单元。我们的核心洞察在于：像素决定图像如何被渲染，而图层决定图像如何被创建、理解和编辑。正如人类设计师通过图层而非原始像素来创建和操作视觉内容，UniWorld-Design 为多模态生成模型赋予了图层原生的设计空间。UniWorld-Design 包含两个模型。Text-to-RGBA（T2RGBA）模型直接从文本生成独立的 RGBA 素材；Image-to-Layer（I2L）模型以一张已完成的图像、一条全局指令和逐层提示为条件，联合生成有序且完整的语义 RGBA 图层。其指令接口支持顶层分解、递归分解和定向提取，使分层成为面向智能体编辑的指令可寻址操作。由于 I2L 学习的是完整的语义对象而非可见像素的分割，其图层在移动或移除后仍然可用。在 Crello 基准上，与 Qwen-Image-Layered 相比，I2L 将逐层 RGB L1 误差降低了 37%，并在 Alpha Soft IoU 上取得了 34% 的相对提升。此外，T2RGBA 取得了最高的 CLIP Score，优于 LayerDiffuse 和 OmniAlpha。

English

We introduce UniWorld-Design, a framework that redefines image generation from flat pixel synthesis to structured visual composition, with semantic RGBA layers as the atomic units of generation, understanding, and editing. Our key insight is that pixels define how an image is rendered, whereas layers define how an image is created, understood, and edited. Just as human designers create and manipulate visual content through layers rather than raw pixels, UniWorld-Design equips multimodal generative models with a layer-native design space. UniWorld-Design comprises two models. The Text-to-RGBA (T2RGBA) model generates standalone RGBA assets directly from text. The Image-to-Layer (I2L) model conditions on a finished image, a global instruction and per-layer prompts, and jointly produces ordered, complete semantic RGBA layers. Its instruction interface supports top-level decomposition, recursive decomposition and targeted extraction, making layering an instruction-addressable operation for agentic editing. Because I2L learns complete semantic objects rather than visible-pixel partitions, its layers stay usable when moved or removed. On the Crello benchmark, I2L reduces per-layer RGB L1 error by 37% and achieves a 34% relative improvement in Alpha Soft IoU over Qwen-Image-Layered. Separately, T2RGBA achieves the highest CLIP Score, outperforming LayerDiffuse and OmniAlpha.