Spatial Action Review: A Visual Analytics Dashboard for Auditing Language-to-Action Hand-offs in Electron Microscopy

📄 arXiv: 2609.24470v1 📥 PDF

作者: Samia Mohinta, Albert Cardona

分类: cs.CV

发布日期: 2026-09-21

备注: Accepted at the IEEE VIS 2026 Workshop on Visual Analytics in the Age of Autonomous Science (VAxAutoSci)


💡 一句话要点

提出Spatial Action Review以解决电子显微镜中的语言到动作交接问题

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 电子显微镜 多模态大语言模型 视觉问答 可视分析 科学图像分析 人机协作 审计工具

📋 核心要点

  1. 现有方法在科学图像分析中,语言到动作的交接存在无声失败问题,导致下游任务的可靠性降低。
  2. 本文提出Spatial Action Review,通过可视化审计工具来连接语言答案与空间动作,帮助识别和修正交接中的错误。
  3. 实验表明,在541个图像区域中,54.4%的记录存在动作不可靠的情况,显示出该方法在审计过程中的有效性。

📝 摘要(中文)

多模态大语言模型(MLLMs)作为科学图像分析的接口日益受到关注,尤其是在视觉问答(VQA)与空间输出的结合中。本文提出了Spatial Action Review,一个用于审计电子显微镜(EM)线粒体分析中语言到动作交接失败的可视分析仪表板。该工具通过答案-动作账本、任务-数据集风险图和图像区域审计视图连接答案与动作,揭示了答案正确但动作不可靠的情况。实验结果显示,在541个图像区域中,54.4%的记录存在这种无声失败,且答案的正确性与动作可靠性之间的关联较弱。

🔬 方法详解

问题定义:本文旨在解决电子显微镜分析中,语言答案与后续空间动作之间的无声失败问题。现有方法在审计过程中未能有效识别这些错误,导致下游任务的可靠性受到影响。

核心思路:Spatial Action Review通过可视化仪表板,将语言答案与对应的空间动作进行关联,提供审计和反馈机制,以便于监督者对动作的可靠性进行评估和修正。

技术框架:该方法包括三个主要模块:答案-动作账本,用于记录答案与动作的对应关系;任务-数据集风险图,展示不同任务在数据集中的风险分布;图像区域审计视图,允许用户对特定图像区域进行详细审计。

关键创新:该研究的创新点在于将语言答案与空间动作的审计过程可视化,揭示了答案正确性与动作可靠性之间的弱关联性,填补了现有方法在这一领域的空白。

关键设计:在设计中,采用了可调节的动作可靠性门限,以支持重新审计,并记录监督者的决策(接受、升级、保持或标记修订),确保审计过程的灵活性和准确性。

🖼️ 关键图片

fig_0
fig_1

📊 实验亮点

实验结果显示,在541个图像区域中,54.4%的记录存在动作不可靠的情况,且27.4%的记录为无声失败。答案的正确性与动作可靠性之间的关联较弱,仅提高5.8个百分点,表明该方法在识别和审计交接错误方面的有效性。

🎯 应用场景

Spatial Action Review在电子显微镜分析中具有广泛的应用潜力,能够提高科学研究的自动化程度和结果的可靠性。该工具不仅适用于生物医学领域,还可以扩展到其他需要图像分析的科学研究中,促进人机协作的效率提升。

📄 摘要(原文)

Multimodal large language models (MLLMs) are increasingly explored as interfaces for scientific image analysis, where a visual question-answering (VQA) response may be paired with a spatial output that guides a downstream stage. A supervisor reads the language answer, while a downstream workflow such as segmentation or region review consumes the point-set output. We call this transition from inspecting the answer to relying on its point action the language-to-action hand-off. A silent failure occurs when the answer is correct while the paired action misses annotated objects needed downstream, so answer-based oversight clears a region whose action is unreliable. We introduce Spatial Action Review, a visual analytics dashboard for auditing this failure mode in electron microscopy (EM) mitochondria analysis. It links paired answer-action records through an answer-action ledger, a task-by-dataset risk map, and an image-region audit view, connecting aggregate patterns to image evidence while an adjustable action-reliability gate supports re-audit. The review ends in a human-AI hand-off, where a supervisor records whether the action is accepted, escalated, held under a stricter gate, or flagged for model revision. Across 541 image regions from an EM-adapted Qwen3-VL case-study run, point actions fail the gate in 54.4% of records with a correct VQA response, and 27.4% of all records are silent failures. A correct answer is associated with only a 5.8-percentage-point higher probability of a reliable action, with a bootstrap interval spanning zero; the point-biserial correlation between answer correctness and object coverage is 0.061. This weak coupling persists across five model conditions on 753 matched image regions. Spatial Action Review makes answer-action mismatches visible and ties them to image evidence and a recorded decision before MLLM outputs enter autonomous scientific workflows.