Orthopantomogram (OPG) Image Analysis Using Bounding Box Algorithm

Abstract

The practice of AI in dentistry can increase dentist precision and save time. It can also provide a second opinion and confidence to the patient [1]. The system proposed in this paper works on ML, cloud computing, and python, which are trending tech in modern times. It promises to provide better performance with accurate results due to faster R-CNN implemented bounding boxes and the scope of batching and processing multiple OPG images in the cloud, real-time learning of ML model, and delivery of diagnosis report securely by web, mobile, and desktop applications. Even patients can donate their OPG images without their personal information to elevate the mastery of the ML model. This paper contains a section on methodology, result analysis, and conclusion. The methodology section speaks about the working of the model from the input of the raw OPG images, then the processing of the images with help of faster RCNN along with the bounding box. The result analysis describes machine learning model and presents processed images at each step. The initial image is a raw OPG and the final image is annotated with the bounding boxes to represent the defected teeth. The final section describes future scope of the use of this technology for further development.