RCT-PEARL: Annotated periapical radiograph dataset for AI-based quality assessment of root canal treatment

This dataset, named “RCT-PEARL: Annotated periapical radiograph dataset for AI-based quality assessment of root canal treatment” comprises a collection of anonymized intraoral periapical radiographs curated for root canal treatment evaluation. The dataset includes the following: IOPA Radiographs: Radiographic images containing root canal treated teeth. Treated Tooth Annotations: Polygon annotations identifying root canal treated teeth. Root Canal Filling Annotations: Instance-level polygon annotations for individual root canal fillings. Root Apex Keypoints: Keypoint annotations marking the apex locations of treated roots. Filling Quality Labels: Labels for each filling instance, including filling length, lateral seal, voids, and irregularities. Additional Findings: Annotations for separated instruments and missed canals where visible. This dataset aims to support the development of multi-task machine learning pipelines for automated root canal treatment quality assessment, including treated tooth localization, root filling segmentation, apex detection, and filling-level quality evaluation.