MICON-Bench: Benchmarking and Enhancing Multi-Image Context Image Generation in Unified Multimodal ModelsAuthors:Mingrui Wu, Hang Liu, Jiayi Ji, Xiaoshuai Sun, Rongrong JiDeep-Dive Summary:这份学术论文介绍了一个名为MICON-Bench的基准测试旨在评估和增强统一多模态模型UMMs在多图上下文生成Multi-Image Context Generation方面的能力。| :—: | :—: | :—: | :—: || BAGEL | 0.3586 | 0.9155 | 0.8766 | 0.6073 ||BAGEL DAR|0.3612|0.9201|0.8828|0.6018|| OmniGen2 | 0.3646 | 0.9102 | 0.8742 | 0.6373 ||OmniGen2 DAR|0.3648|0.9130|0.8757|0.6327|表 5参考图像数量对 UMMs 性能的影响。ModelRef2Ref3Ref4Ref5BAGEL88.5084.3775.1166.36OmniGen292.1889.5274.9267.00图 4DAR 抑制噪声注意力并重新聚焦目标。红色框代表被抑制的无关区域绿色框代表被增强的目标区域。6. 结论 (Conclusion)MICON-Bench 为多图上下文生成提供了严谨的评估平台而 DAR 机制则提供了一种有效且低成本的手段来解决当前 UMMs 在跨图像推理中的幻觉问题。这两者共同为开发更可靠的多模态生成系统奠定了基础。Original Abstract:Recent advancements in Unified Multimodal Models (UMMs) have enabled remarkable image understanding and generation capabilities. However, while models like Gemini-2.5-Flash-Image show emerging abilities to reason over multiple related images, existing benchmarks rarely address the challenges of multi-image context generation, focusing mainly on text-to-image or single-image editing tasks. In this work, we introduce \textbf{MICON-Bench}, a comprehensive benchmark covering six tasks that evaluate cross-image composition, contextual reasoning, and identity preservation. We further propose an MLLM-driven Evaluation-by-Checkpoint framework for automatic verification of semantic and visual consistency, where multimodal large language model (MLLM) serves as a verifier. Additionally, we present \textbf{Dynamic Attention Rebalancing (DAR)}, a training-free, plug-and-play mechanism that dynamically adjusts attention during inference to enhance coherence and reduce hallucinations. Extensive experiments on various state-of-the-art open-source models demonstrate both the rigor of MICON-Bench in exposing multi-image reasoning challenges and the efficacy of DAR in improving generation quality and cross-image coherence. Github: https://github.com/Angusliuuu/MICON-Bench.PDF Link:2602.19497v1部分平台可能图片显示异常请以我的博客内容为准