Precise Dental Staging Method through Panoramic Radiographs Based on Deep Learning
Abstract
In this paper, we use a deep learning method to realize automatic dental staging, which is an image classification task. Sample imbalance is found to influence the staging accuracy of deep learning method. Aiming at the problem, an optimized method for dental staging is proposed. Firstly, based on the thought of coarse-to-fine, the dental staging task is converted into a binary-classification and two multi-classification tasks to reduce the difficulty of each classification task. Secondly, data augmentation is used in multi-classification to balance the samples in different categories. Besides, the Easy-Ensemble method is used to reduce the influence of sample imbalance in the binary-classification task. According to the experiment results, the dental staging effect of the proposed method is improved in accuracy, recall, and accuracy compared with the unoptimized method, which verifies the availability of the proposed method.
Add the publication’s full text or supplementary notes here. You can use rich formatting such as including code, math, and images.
HAN Hongcheng
PhD Student in Control Science and Engineering
My research interests focus on AI-assisted medical image analysis.