arXiv cs.AI· Zhiqiang Wu, Shuang Sun, Jiale Zhang, Jing Li, Yibo Chen, Xu Tang, Yao Hu·· 3 小时前AI 评分33
DeCoRL：面向多图组合一致性的分解式强化学习方法
DeCoRL: Decomposed Consistency Reinforcement Learning for Multi-Image Composition
AI 导读
研究者提出 DeCoRL 强化学习框架，将多图组合（MICo）的整体一致性分解为对象级一致性，先建立参考图之间的指令相关对象对应关系，再用基于 VLM 的奖励模型生成对象级反馈以指导 RL 后训练。该工作还引入覆盖 2 至 6 张参考图的 DeCo-Bench 基准，在 DeCo-Bench 和公开的 MICo-Bench 上均取得一致提升。
正文 · 原文
Abstract:Multi-image composition (MICo) is a critical yet challenging capability for large-scale image editing models, where consistency degrades rapidly as the number of reference images increases. While supervised fine-tuning (SFT) can equip models with basic MICo ability, it provides only implicit supervision on the composed image and often fails to learn reliable object-level correspondence across multiple references. We propose DeCoRL, a reinforcement-learning (RL) framework that improves MICo consistency by decomposing holistic MICo consistency into object-level consistency. DeCoRL first establishes instruction-relevant object correspondences across reference images and localizes the corresponding regions in the generated image. It then uses a VLM-based reward model trained on single-object consistency annotations to produce fine-grained, object-wise feedback, which is aggregated into a multi-object consistency reward to guide RL post-training. We further introduce DeCo-Bench, a benchmark covering compositions from 2 to 6 reference images. Experiments on DeCo-Bench and the public MICo-Bench demonstrate consistent improvements in MICo consistency.
| Comments: | NeurIPS |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.39265 [cs.CV] |
| (or arXiv:2610.39265v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.39265 arXiv-issued DOI via DataCite |
Submission history
From: Zhiqiang Wu [view email]
[v1]
Thu, 29 Oct 2026 03:18:03 UTC (5,410 KB)
来源：arXiv cs.AI · arxiv.org