Artificial intelligence based on Convolutional Neural Network for detecting dental caries on bitewing and periapical radiographs

Objectives: This narrative review is written to describe the accuracy of caries detection and find out clinical implications future prospects using Convolutional Neural Network (CNN) determine radio-diagnosis dental in bitewing periapical radiographs.
 Review: The databases used for literature searching this were PubMed, Google Scholar, Science Direct. inclusion criteria original article, case report, textbook English Bahasa Indonesia, published within 2011-2021. exclusion articles that full text could not be accessed, research article did provide methods used, duplication articles. In review, a total 33 literatures consisting 30 three textbooks reviewed, including four on CNN detection.
 Conclusion: Results reveal GoogLeNet produces best compared Fully (FCN) U-Net radiographs. Nonetheless, positive predictive value (PPV), recall, negative (NPV), specificity, F1-score, values these architectures indicate good performance. differences each CNN’s performances detect are determined by number trained datasets, architecture’s layers, complexity architectures. conclusion can as an alternative caries, increasing diagnostic time efficiency well preventing errors due dentist fatigue. Yet able substitute expertise radiologist. Therefore, it need revalidated radiologist avoid errors.