"Oral Mucosa Capillary Dermoscopy Dataset with Expert Annotations for Segmentation Analysis"

"This dataset contains 212 high-resolution oral mucosal dermoscopy images (1440\u00d71080 pixels) from 106 subjects with expert-annotated capillary segmentation masks, representing the first publicly available dataset of its kind for oral microvascular analysis. Images were captured using polarized dermoscopy at 20\u00d7 magnification at Changsha Stomatological Hospital, covering healthy controls, oral submucous fibrosis patients, and other oral mucosal conditions. Each image includes pixel-level capillary annotations validated through triple-expert consensus with 0.875 inter-observer Dice agreement, enabling reproducible research in automated capillary segmentation, zero-shot learning methods, and computer-aided diagnosis for oral mucosal diseases."