arXiv cs.CL· Lijia Wu, Zheng Li, Yuxuan Liu, Jingwen Zhang, Penghai Zhao, Xiang Li·· 12 小时前AI 评分31
SciFigBench：科学方法图理解基准发布
SciFigBench: A Benchmark for Scientific Method Figure Understanding
AI 导读
SciFigBench 是面向科学方法图理解的基准，从真实论文中整理 4,352 张方法图和 40,670 个 QA 对，覆盖密集内容感知、视觉关系理解与信息流追踪三项能力。配套提出免训练视觉提示模块 SciFigCue，将箭头依赖外化为视觉关系证据。实验显示开源与闭源 VLM 在视觉关系理解上仍明显受限，SciFigCue 无需更新参数即可在多项任务上提升表现。
正文 · 原文
Abstract:Scientific method figures often provide researchers with a clear visual entry point into a paper's methodology and core contributions, yet their dense layouts and complex information flows are difficult for Vision Language Models (VLMs) to interpret. Existing benchmarks predominantly focus on coarse-grained tasks, such as context-based retrieval or holistic comprehension, overlooking the fine-grained content and inter-module relations within complex methodological diagrams. To address these overlooked aspects, we introduce Scientific Figure Benchmark (SciFigBench), a benchmark for evaluating core capabilities required for scientific method figure understanding. Specifically, it targets three capabilities: dense content perception, visual relation comprehension, and information-flow tracing. For this purpose, we curate 4,352 scientific method figures from real-world research papers and construct 40,670 QA pairs to operationalize these capabilities as measurable question-answering tasks. All annotations are generated by dedicated pipelines and then manually verified. We further propose SciFigCue, a training-free visual prompting module that externalizes arrow-induced dependencies as visual relational evidence for method-figure reasoning. Experiments across open-source and proprietary VLMs show that even strong models remain substantially limited in visual relation comprehension, while SciFigCue improves performance across several tasks without parameter updates. The code and related resources are available at this https URL.
| Comments: | Accepted by EMNLP 2026 Main Conference. Code and benchmark resources: this https URL |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.27087 [cs.CV] |
| (or arXiv:2610.27087v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27087 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Lijia Wu [view email]
[v1]
Mon, 19 Oct 2026 18:00:07 UTC (4,066 KB)
来源：arXiv cs.CL · arxiv.org