Speaker
Description
Digital dentistry refers to the digitalization of the entire process of dental treatment, including bone measurements, virtual surgery planning, occlusion analysis, and implant placement. With the recent integration of deep learning technologies, digital dentistry has become increasingly accelerated, and it is expected to significantly improve the efficiency of dentists' work and enhance patient satisfaction. In this presentation, we will introduce an integrated core software developed for digital dentistry. This software has the ability to improve the image quality of cone-beam computed tomography (CBCT) and automatically segment the skull and individual teeth. It can also register all data obtained from various sources. Specifically, the metal artifacts in CBCT images are corrected using a deep learning-based correction algorithm. We have adopted an iterative correction method that imposes data fidelity outside the metal trace to retain the anatomical structures of the target image. Considering the difficulty of obtaining a paired dataset of metal-free and metal-affected CT scans, the simulation-based training method for metal artifact reduction was adopted. Moreover, the software uses a fully automated method to identify and segment individual teeth from dental CBCT images. The upper and lower jaws panoramic images was generated and used to overcome the computational complexity caused by high-dimensional data and the curse of dimensionality associated with limited training dataset. The integration of intra-oral scan (IOS) and dental CBCT images into one image is performed in three steps: individual tooth segmentation and identification modules for IOS and CBCT data, and global-to-local tooth registration between IOS and CBCT. Finally, a fully automatic registration method of dental CBCT and face scan data is adopted. This involves generating surfaces from the measured CT and facial scans, detecting facial points corresponding to the 3D landmarks, estimating 3D landmark positions, and performing the initial registration. The Iterative Closest Point method is then applied to improve the accuracy and efficiency of the registration of CT and face surfaces. Overall, these technologies are essential elements of digital dentistry that can significantly improve the accuracy and efficiency of dental treatment.
References
[1] Park, H. S., Jeon, K., Lee, S. H., & Seo, J. K. (2022). Unpaired-Paired Learning for Shading Correction in Cone-Beam Computed Tomography. IEEE Access, 10, 26140-26148.
[2] Park, H. S., Seo, J. K., Hyun, C. M., Lee, S. M., & Jeon, K. (2022). A fidelity‐embedded learning for metal artifact reduction in dental CBCT. Medical Physics, 49(8), 5195-5205.
[3] Jang, T. J., Kim, K. C., Cho, H. C., & Seo, J. K. (2022). A fully automated method for 3D individual tooth identification and segmentation in dental CBCT. IEEE transactions on pattern analysis and machine intelligence, 44(10), 6562-6568.
[4] Jang, T. J., Yun, H. S., Kim, J. E., Lee, S. H., & Seo, J. K. Fully automatic integration of dental CBCT images and full-arch intraoral impressions with stitching error correction via individual tooth segmentation and identification, arXiv:2112.01784
| Keywords | Digital dentistry, deep learning |
|---|
