The Development of Dental Problem Detection Device Using Inception V3 and Data Augmentation

This study focuses on the development of a dental problem detection device using the Inception V3 deep learning model and advanced data augmentation techniques. Dental problems such as cavities, impacted wisdom teeth, and crowded teeth pose significant health risks, often requiring advanced imaging for accurate diagnosis. The proposed device utilizes an intraoral camera to capture digital tooth images, which are then processed using the Inception V3 model to identify three dental problems- dental caries, crowded teeth, and partially erupted wisdom teeth. By employing image data augmentation and leveraging datasets from various dental repositories, the study seeks to enhance the accuracy and reliability of dental diagnostics. As a result, the model achieved a total accuracy of 87.2%, in which the model attained a precision of 92%, a recall of 93%, and an Fl-score of 92.5% for caries detection. In detecting crowded teeth, it recorded a precision of 81.1 %, a recall of 90.9%, and an Fl-score of 85.7%. For partially erupted wisdom teeth detection, the precision was 88.8%, the recall 78%, and the Fl-score 83 %. These results indicate that while the model demonstrates strong overall performance, further improvements are necessary to improve the detection of crowded teeth and partially erupted wisdom teeth. This innovative approach endeavors to provide a non-invasive, efficient, and precise tool for dentists, potentially revolutionizing dental healthcare through the integration of advanced AI technologies into routine dental practices.