Design and Comparison of Transfer Learning for Dental Caries Detection

Dental caries is one of the most serious dental health issues affecting people of all ages. It is an infectious disease that degrades tooth structure because of dental cavities. Due to constraints such as the enormous amount of money and time required, Artificial Intelligence (AI) has been used in medical research in the field of oral healthcare. Deep learning (DL), an AI branch, is currently a growing field that is widely used in dentistry. DL is a solid foundation for dentists to give better, more efficient care to patients while also saving time. AI supports dentists in meeting patients' expectations while also ensuring excellent treatment and better oral health care. Furthermore, DL can predict clinical case failures and propose dependable therapies. This strategy aids in lowering morbidity rates and improving the quality of dental care in the population. Using transfer learning techniques, this study aims to automate the detection of dental cavities. The two transfer learning algorithms used are ResNet50 and MobileNet. 395 normal and 395 caries dental images from Kaggle are used to develop the transfer learning model. The optimal model is determined by examining both transfer learning models. The MobileNet gives the highest accuracy when compared with ResNet50. The accuracy rate achieved by the MobileNet model is 96.12%.