3D System and Segmentation for Dental Care

Modern dental imagery has undergone significant changes as a result of technological advancements. However, for an accurate diagnosis, multiple imaging methods are required to ensure improved efficiency. This makes the application of dental imagery for predicting treatment, detailed diagnosis, and tooth segmentation laborious, computationally intensive, and expensive. Moreover, artifacts introduced by image degradation prevent accurate analysis and are also considered one of the fundamental problems in dental imagery.

To overcome the common issue of spatial resolution in dental imagery and the laborious process of tooth segmentation, we utilized the advancements in computer vision technologies to build a 3D dental system. We used DICOM format data to extract information in three-dimension and computer vision algorithms such as binarization, edge detection, and contour tracing with clustering for tooth segmentation. Our system outputs a point cloud-based representation and sets the foundation for tooth segmentation analysis.

Presentation Video | Source Code

Contributors: Aditya Bapat, Ajith Potluri, Ziyuan Wei, Monalisa Malani

24678: Computer Vision
Mechanical Engineering Department
Carnegie Mellon University – Pittsburgh
Fall 2022