PRIMA：多图像视觉-语言模型用于推理分割
PRIMA: Multi-Image Vision-Language Models for Reasoning Segmentation
AI总结：
PRIMA通过整合像素级定位与多图像推理，提升了多图像分割任务的性能，其在Recall和S-IoU上分别优于现有基线7.83%和11.25%。
摘要：
尽管在大型视觉-语言模型（LVLMs）的能力上取得了显著进展，但现有像素定位模型仅在单图像设置中运行，限制了它们在多图像之间进行详细、细粒度比较的能力。相反，当前的多图像理解模型缺乏像素级定位。我们的工作通过引入多图像像素定位推理任务，结合PRIMA，一种整合像素级定位与稳健多图像推理的LVLM，以产生上下文丰富、像素定位的解释。PRIMA的核心是SQuARE，一个视觉模块，它在融合到语言骨干之前，将跨图像的相对上下文注入到紧凑的查询基于的视觉标记中。为了支持训练和评估，我们整理了M4SEG，一个新的多图像推理分割基准，包含约744,000个问题-答案对，需要在多个图像上进行细粒度视觉理解。PRIMA在Recall和S-IoU上分别比最先进的基线提高了7.83%和11.25%。消融研究进一步证明了所提出的SQuARE模块在捕捉跨图像关系方面的有效性。
英文摘要：
Despite significant advancements in Large Vision-Language Models (LVLMs)' capabilities, existing pixel-grounding models operate in single-image settings, limiting their ability to perform detailed, fine-grained comparisons across multiple images. Conversely, current multi-image understanding models lack pixel-level grounding. Our work addresses this gap by introducing the task of multi-image pixel-grounded reasoning alongside PRIMA, an LVLM that integrates pixel-level grounding with robust multi-image reasoning to produce contextually rich, pixel-grounded explanations. Central to PRIMA is SQuARE, a vision module that injects cross-image relational context into compact query-based visual tokens before fusing them with the language backbone. To support training and evaluation, we curate M4SEG, a new multi-image reasoning segmentation benchmark consisting of $\sim$744K question-answer pairs that require fine-grained visual understanding across multiple images. PRIMA outperforms state-of-the-art baselines with $7.83\%$ and $11.25\%$ improvements in Recall and S-IoU, respectively. Ablation studies further demonstrate the effectiveness of the proposed SQuARE module in capturing cross-image relationships.
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。
