SinaiU-OMP dataset: A curated histopathological image dataset for oral and maxillofacial pathology to support ai research

Résumé fourni par la source

High-quality, well-annotated datasets are a fundamental prerequisite for the development of reliable artificial intelligence (AI) systems in digital and computational pathology. This paper introduces the SinaiU-OMP dataset, a curated collection of 300 real histopathological images representing five essential diagnostic classes in oral and maxillofacial pathology: Normal oral mucosa, Oral Squamous Cell Carcinoma (OSCC), Odontogenic Keratocyst (OKC), Ameloblastoma–Plexiform (AB), and Pleomorphic Adenoma (PA). All images are derived from routine diagnostic practice, ethically approved for research use, and labeled through expert-confirmed histopathological diagnoses. Each case is accompanied by structured metadata, standardized labeling, and a reproducible directory architecture to support downstream reuse in computational pathology research. The dataset is designed as a foundational, AI-ready resource for model development, algorithm benchmarking, educational applications, and methodological research in oral pathology. A technical validation study using a combined multi-source dataset confirms the internal consistency, structural integrity, and usability of the dataset within standard machine learning pipelines. The SinaiU-OMP dataset is publicly available as an open research resource and provides a standardized reference framework to support reproducible and transparent AI research in oral and maxillofacial pathology.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.