Deep Learning-based Cavity Detection in Diverse Intraoral Images: A Web-based Tool for Accessible Dental Care

Abstract

Dental cavities represent a widespread oral health issue on a global scale, impacting individuals across all age groups. The conventional approach to detecting cavities involves a visual examination by dentists, which is not only time-consuming but also subjective. Current methods for dental cavity detection heavily rely on subjective visual inspections, which may overlook early or concealed cavities. In this research paper, we present CatchCavity, a web application tool that utilizes deep learning for cavity detection in teeth. The diagnostic tool enables users to upload dental images, facilitating the assessment of dental cavity status. Furthermore, the web application functions as an online dental diagnostic service, providing the capability to securely store patients’ dental records and information in a dedicated database. The system is trained on a dataset of annotated images and employs a convolutional neural network (CNN) architecture for accurate cavity detection. We evaluate the system's performance using metrics such as accuracy and loss. Our results showcase that the proposed system attains a high level of accuracy and efficiency in detecting dental cavities, achieving an overall accuracy of 98.7%. Additionally, our system surpasses traditional cavity detection methods, highlighting the possibility of deep learning approaches to enhance oral health outcomes.