English

Decoding Scientific Experimental Images: The SPUR Benchmark for Perception, Understanding, and Reasoning

Computer Vision and Pattern Recognition 2026-05-27 v2 Computational Engineering, Finance, and Science

Abstract

We introduce SPUR, a comprehensive benchmark for scientific experimental image perception, understanding, and reasoning, comprising 4,264 question-answering (QA) pairs derived from 1,084 expert-curated images. SPUR features three key innovations: (1) Panel-Level Fine-Grained Perception: evaluating the visual perception of multimodal large language models (MLLMs) across three dimensions (numerical, morphological, and information localization) on six fine-grained panel types; (2) Cross-Panel Relation Understanding: utilizing complex images with an average of 14.3 panels per sample to evaluate MLLMs' ability to decipher intricate cross-panel relations; (3) Expert-Level Reasoning: assessment of qualitative and quantitative reasoning across five experimental paradigms to determine if models can infer conclusions from evidence as human experts do. Comprehensive evaluation of 20 MLLMs and four multimodal Chain-of-Thought (MCoT) methods reveals that current models fall significantly short of the expert-level requirements for scientific image interpretation, underscoring a critical bottleneck in AI for Science (AI4S) research.

Keywords

Cite

@article{arxiv.2604.27604,
  title  = {Decoding Scientific Experimental Images: The SPUR Benchmark for Perception, Understanding, and Reasoning},
  author = {Junpeng Ding and Zichen Tang and Haihong E and Mengyuan Ji and Yang Liu and Haolin Tian and Haiyang Sun and Pengqi Sun and Yang Xu and Yichen Liu and Haocheng Gao and Zijie Xi and Ruomeng Jiang and Peizhi Zhao and Rongjin Li and Yuanze Li and Jiacheng Liu and Zhongjun Yang and Jintong Chen and Siying Lin},
  journal= {arXiv preprint arXiv:2604.27604},
  year   = {2026}
}

Comments

Accepted to ACL 2026 Main Conference

R2 v1 2026-07-01T12:43:10.982Z