Development of a PCA-Based Post-Processing Algorithm for Individual Teeth Segmentation in Dental X-ray Images

With advancements in artificial intelligence, various fields, including magnetic medical science, are achieving results by utilizing it. In particular, the use of AI based medical image processing in dentistry is predicting oral conditions through image processing such as panoramic images and Computed Tomography (CT) contributing to the improvement of dental services. This study proposes an algorithm to improve the accuracy of individual tooth segmentation based on dental panoramic images to promote precise diagnosis and treatment planning, and to contribute to medical imaging technology with magnetic medical imaging. U-Net is improved to segment images and a post-processing step based on principal component analysis (PCA) was applied. The results show that the proposed post-processing algorithm outperforms the existing method, and the accuracy of tooth segmentation was improved. It is expected to remarkably improve clinical efficiency in dentistry and provide a foundation for the magnetic medical imaging techniques in the diagnosis and treatment of oral diseases.