Enhancing Dental Radiography: Teeth Instance Segmentation in Panoramic X-Rays

Accurate tooth segmentation in panoramic X-ray images holds significant importance in the field of dental healthcare, as it provides vital information for clinical diagnoses, orthodontic treatments, surgical procedures, tooth morphology analysis, and dental implant planning. This study focuses on instance segmentation using panoramic dental X-ray images, aiming to address the need for precise and efficient segmentation of individual teeth within these radiographs. To tackle this challenge, we employ the innovative YOLOv8-Seg model on the novel OdontoAI Open Panoramic Radiographs ( \(O^{2}PR\) ) dataset. Multiple versions of YOLOv8-Seg, including Nano, Small, Medium, Large, and XLarge, were tested, with YOLOv8-SegXLarge emerging as the winner architecture, showcasing the best performance across various metrics. The results demonstrate that YOLOv8-Seg effectively accomplishes tooth instance segmentation in panoramic radiographs, highlighting its potential value in enhancing dental imaging and analysis techniques.