2021 Volume 63 Issue 3 Pages 119-128
Standardization of medical care can be achieved by double-checking through the use of artificial intelligence (AI). Panoramic radiography, which is used daily in dental and oral surgery practice, captures standardized images that can evaluate the condition of the oral and maxillofacial region. This report presents an AI system for detecting furcation involvement through panoramic radiography. We captured 10,640 panoramic radiography images showing furcation involvement as training data and constructed a deep convolution neural network (DCNN) that automatically detects the images.
We used 170 images revealing furcation involvement in the mandibular molars, and the accuracy, sensitivity, specificity, precision, recall, and F-score were evaluated from those images. Our data showed an accuracy, sensitivity, specificity, precision, recall and F-score of 96.4%, 95.6%, 97.1%, 96.3%, 95.6%, and 0.96 respectively. We developed an AI system for detecting furcation involvement through panoramic radiography using DCNN deep learning. The proposed system is expected to be useful to both dental professionals and patients. These findings demonstrate that our AI system can improve clinical diagnosis in the future.
