arXiv 2603.086522026-08-07cs.AI

CoCo：用于文本到图像预览和稀有概念生成的代码作为CoT

CoCo: Code as CoT for Text-to-Image Preview and Rare Concept Generation

South China University of Technology（南方科技大学）；StepFun；Institute of Automation Chinese Academy of Sciences（中国科学院自动化研究所）；Nanyang Technological University（南洋理工大学）；The Chinese University of Hong Kong（香港中文大学）

Haodong Li, Chunmei Qing, Huanyu Zhang, Dongzhi Jiang, Yihang Zou, Hongbo Peng, Dingming Li, Yuhong Dai, ZePeng Lin, Juanxi Tian, Yi Zhou, Siqi Dai, Jingwei Wu, Pheng-Ann Heng

备注：: 21 pages, 7 figures, and 3 tables. Accepted to ECCV 2026

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

CoCo通过代码驱动的推理框架，提升文本到图像生成的精度和可控性，实现结构化图像生成和稀有概念生成。

摘要：

最近统一多模态模型（UMMs）的进步显著推动了文本到图像（T2I）生成，特别是在通过链式思维（CoT）推理的整合方面。然而，现有的基于CoT的T2I方法大多依赖于抽象的自然语言规划，这缺乏复杂空间布局、结构化视觉元素和密集文本内容所需的精度。在本工作中，我们提出了CoCo（代码作为CoT），一种以代码驱动的推理框架，将推理过程表示为可执行代码，从而实现图像生成的明确且可验证的中间规划。给定一个文本提示，CoCo首先生成指定场景结构布局的可执行代码，然后在沙盒环境中执行以渲染一个确定性的草稿图像。模型随后通过细粒度图像编辑来完善此草稿，以产生最终的高质量结果。为了支持这种训练范式，我们构建了CoCo-10K，一个经过精心编排的数据集，包含结构化草稿-最终图像对，旨在教授结构化草稿的构建和纠正视觉细化。在StructT2IBench、OneIG-Bench和LongText-Bench上的实证评估表明，CoCo在直接生成上实现了+68.83%、+54.8%和+41.23%的改进，同时在其他由CoT赋能的生成方法中也表现更优。这些结果表明，可执行代码是一种有效的且可靠的推理范式，能够实现精确、可控和结构化的文本到图像生成。代码可在：https://github.com/micky-li-hd/CoCo获取。

英文摘要：

Recent advancements in Unified Multimodal Models (UMMs) have significantly advanced text-to-image (T2I) generation, particularly through the integration of Chain-of-Thought (CoT) reasoning. However, existing CoT-based T2I methods largely rely on abstract natural-language planning, which lacks the precision required for complex spatial layouts, structured visual elements, and dense textual content. In this work, we propose CoCo (Code-as-CoT), a code-driven reasoning framework that represents the reasoning process as executable code, enabling explicit and verifiable intermediate planning for image generation. Given a text prompt, CoCo first generates executable code that specifies the structural layout of the scene. The code is then executed in a sandboxed environment to render a deterministic draft image. Subsequently, the model refines this draft through fine-grained image editing to produce the final high-fidelity result. To support this training paradigm, we construct CoCo-10K, a curated dataset containing structured draft-final image pairs designed to teach both structured draft construction and corrective visual refinement. Empirical evaluations on StructT2IBench, OneIG-Bench, and LongText-Bench show that CoCo achieves improvements of 68.83%, 54.8%, and 41.23%, respectively, over direct generation, while also outperforming other CoT-enhanced generation methods. These results demonstrate that executable code is an effective and reliable reasoning paradigm for precise, controllable, and structured text-to-image generation.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。