Advanced YOLOv7-Based Detection and Enumeration of Teeth and Restorations in Panoramic X-Rays

Dental radiography plays a crucial role in oral health assessment, aiding dentists in diagnosis and treatment planning. This study introduces an automated method for enumerating teeth, implants, and pontics of dental bridges in panoramic dental X-rays using the YOLOv7 network. The method detects these elements and employs an innovative optimization and dental bridge segmentation algorithm for accurately enumerating teeth and restorations. In this study, we analyzed 1000 panoramic dental X-ray images, including 450 with dental bridges and 64 with implants. For evaluation, 200 images were selected, revealing a mean Average Precision (mAP) of 98.15% for tooth and restoration detection, with overall precision of 0.9774 and overall recall of 0.9803 for enumeration, demonstrating practical utility of the proposed method in clinical dentistry.