Unlocking the Visual Record of Materials Science: A Large-Scale Multimodal Dataset from Scientific Literature

Hugging Face Daily Papers Papers

Summary

Introduces MatMMExtract pipeline to decompose compound scientific figures into panels and annotate them using LLMs, creating the MatSciFig dataset of over 390,000 image-text pairs for vision-language learning in materials science.

The materials science literature encodes decades of experimental knowledge in figures, yet this visual record remains locked away and inaccessible to AI at scale. The core difficulty is structural: most scientific figures are compound, with a single caption describing multiple sub-panels simultaneously, making direct image-text pairing unreliable. We present MatMMExtract, an end-to-end open-source pipeline that resolves this by decomposing compound figures into individual sub-panels and generating structured, grounded annotations using a large language model guided by a curated materials science taxonomy. Applied to 14,810 open-access articles, MatMMExtract produces MatSciFig; 391,606 panel-level image-text pairs from 180,571 figures, each annotated with a sub-caption, a two-level visualisation category spanning 19 classes and over 100 subtypes, and a scientific summary. To enable accurate panel localisation, we introduce MaterialScope, a domain-specific detection dataset of 2,811 manually annotated materials science figures, on which a fine-tuned YOLO12-m detector achieves mAP_50 of 0.9227. Among six benchmarked language models, Gemini 3.1 Flash Lite delivers the best cost-quality trade-off for annotation generation, with 82% of outputs rated good and a hallucination rate of 4.8%. A dual-encoder retrieval baseline on MatSciFig achieves a 4.4 times improvement in R@1 over zero-shot CLIP, demonstrating the dataset's immediate utility for vision-language learning. All resources are released openly to the community.
Original Article
View Cached Full Text

Cached at: 07/01/26, 03:43 PM

Paper page - Unlocking the Visual Record of Materials Science: A Large-Scale Multimodal Dataset from Scientific Literature

Source: https://huggingface.co/papers/2606.29667

Abstract

A novel pipeline called MatMMExtract is introduced that processes compound scientific figures into individual panels and generates structured annotations using large language models, creating a comprehensive dataset for vision-language learning in materials science.

The materials science literature encodes decades of experimental knowledge in figures, yet this visual record remains locked away and inaccessible to AI at scale. The core difficulty is structural: most scientific figures are compound, with a single caption describing multiple sub-panels simultaneously, making direct image-text pairing unreliable. We present MatMMExtract, an end-to-end open-source pipeline that resolves this by decomposing compound figures into individual sub-panels and generating structured, grounded annotations using alarge language modelguided by a curatedmaterials science taxonomy. Applied to 14,810 open-access articles, MatMMExtract produces MatSciFig; 391,606 panel-levelimage-text pairsfrom 180,571 figures, each annotated with a sub-caption, a two-level visualisation category spanning 19 classes and over 100 subtypes, and a scientific summary. To enable accurate panel localisation, we introduce MaterialScope, a domain-specific detection dataset of 2,811 manually annotated materials science figures, on which a fine-tunedYOLO12-m detectorachieves mAP_50 of 0.9227. Among six benchmarked language models, Gemini 3.1 Flash Lite delivers the best cost-quality trade-off for annotation generation, with 82% of outputs rated good and a hallucination rate of 4.8%. Adual-encoder retrievalbaseline on MatSciFig achieves a 4.4 times improvement in R@1 over zero-shotCLIP, demonstrating the dataset’s immediate utility forvision-language learning. All resources are released openly to the community.

View arXiv pageView PDFGitHub0Add to collection

Get this paper in your agent:

hf papers read 2606\.29667

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.29667 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.29667 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.29667 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

MAOAM: Unified Object and Material Selection with Vision-Language Models

Hugging Face Daily Papers

This paper presents MAOAM, a unified vision-language model framework that enables precise object and material selection through text or click interactions for interactive image editing. It introduces a scalable data generation pipeline and shows emergent improvement when combining text and clicks at inference.

Miller-Index-Based Latent Crystallographic Fracture Plane Reasoning with Vision-Language Models

arXiv cs.LG

This paper investigates whether multimodal large language models (MLLMs) can leverage Miller indices as a latent representation to reason about crystallographic fracture geometry from visual inputs, evaluating their ability to infer physically valid plane hypotheses and determine when such representation is applicable across materials like ceramics, glass, metals, and concrete.