BSE Thin-Section Dataset for Petrographic Rock Classification

This dataset contains backscattered electron (BSE) thin-section imagery features used for petrographic rock classification across five mineralogical classes. It was curated for evaluating machine learning and deep learning models, including hybrid stacking ensembles. The dataset accompanies the paper “Hybrid Stacking Ensemble for Petrographic Rock Classification” and is intended as a benchmark for reproducible research.