Detecting dental caries with convolutional neural networks using color images

Oral healthcare systems have garnered significant attention due to the potential for preventive care to reduce both the cost and severity of treatments. Numerous studies have shown promising results using X-ray films for dental disease detection. However, RGB images are rarely utilized in this context. This study aims to perform object detection on specific dental abnormalities using a Convolutional Neural Network (CNN) model. Various models of YOLOv۸ were employed to compare their efficacy, with the dataset collected by a dental specialist, using an oral camera. Our results demonstrate that the YOLOv۸s model achieved a precision of ۸۴%, a recall of ۷۹%, and an mAP@۰.۵ of ۸۵%. This paper highlights the potential of using color images to develop a mobile oral healthcare system to detect dental caries and abnormalities in the future.

کلیدواژه ها:

teeth detection ، oral healthcare ، Yolov۸ ، dental calculus ، convolutional neural network ، deep learning