Sci-ImageMiner: A Multimodal Expert-annotated Dataset of Scientific Figures From Atomic Layer Deposition and Etching Literature

Sci-ImageMiner is a scientific figure understanding dataset built from research papers on Atomic Layer Deposition (ALD) and Atomic Layer Etching (ALE). It contains 1,951 figures from 205 research papers, covering a wide range of scientific figures, including plots, diagrams, images, and tables. The dataset includes expert annotations for figure classification, data extraction, figure summarization, and visual question answering. It is designed to support the development and evaluation of AI models, particularly large vision-language models, that can understand and interpret figures in scientific literature.