A deep-learning artificial intelligence system for estimating chronological age using panoramic radiography in the Korean population

Abstract

The purpose of this study was to suggest a hybrid method based on ResNet50 and ViT in an age estimation model using panoramic radiographs for learning by considering both local features and global information, which is important in estimating age. Transverse and longitudinal panoramic images of 9663 patients were selected and used (4774 males and 4889 females with a mean age of 39 years and 3 months). To compare ResNet50, ViT, and the hybrid model, the MAE, mean square error (MSE), root mean square error (RMSE), and coefficient of determination (R2) were used as metrics. The results confirmed that the age estimation model designed using the hybrid method performed better than those using only ResNet50 or ViT. In addition, when examining the basis for age determination in the hybrid model through attention rollout, it was evident that the proposed model used logical and important factors rather than relying on unclear elements as the basis for age determination.