Automatic design and 3D segmentation of mandible bone using CNN algorithm via exclusive GUI

Nowadays, scientists are looking to decrease dental faults by presenting new approaches. It is obvious that comprehensive information about the anatomic position of the inferior alveolar neural canal is essential to have the most ideal mandible surgery or systemic tooth implant. Accordingly, we present a new approach in this article that can be used to have 3D segmentation and recognition of the mentioned canal in mandible by CBCT image. This approach includes two main steps. In the first step, we train a full convolutional 3D net (FCN) to reach the ability of section recognition, which can recognize the relevant area of the mandible bone. In the next step, we define a 3D U-net, which is similar to FCN, to segment the inferior Alveolar neural (IAN) canal from the lower jaw. Evaluated on publicly available datasets, our method achieved an average Dice coefficient of 86.61%.