A pilot study of a deep learning approach to submerged primary tooth classification and detection.

• DOI: 10.3290/j.ijcd.b994539
• Publication Date: 2021-02-26
• Authors: Seçil ÇALIŞKAN; Nuray Tüloğlu; Özer Çelik; Canan ÖZDEMİR; Sena Kızılaslan; Şule Bayrak
• Keywords: Contextual image classification

Aim The aim of the study was to compare success and reliability an artificial intelligence (AI) application in detection classification submerged teeth panoramic radiographs. Materials methods Convolutional neural network (CNN) algorithms were used detect classify molars. module, based on stateof- the-art Faster R-CNN architecture, processed a radiograph define boundaries A separate testing set evaluate diagnostic performance system it with that experts field. Result rate identification high when evaluated according reference standard. extremely accurate its comparison observers. Conclusions proposed computeraided diagnosis solution is comparable experts. It useful diagnose molars AI prevent errors. In addition, this will facilitate diagnoses pediatric dentists.