Automatic mandibular canal detection using a deep convolutional neural network

Abstract

The practicability of deep learning techniques has been demonstrated by their successful implementation in varied fields, including diagnostic imaging for clinicians. In accordance with the increasing demands in the healthcare industry, techniques for automatic prediction and detection are being widely researched. Particularly in dentistry, for various reasons, automated mandibular canal detection has become highly desirable. The positioning of the inferior alveolar nerve (IAN), which is one of the major structures in the mandible, is crucial to prevent nerve injury during surgical procedures. However, automatic segmentation using Cone beam computed tomography (CBCT) poses certain difficulties, such as the complex appearance of the human skull, limited number of datasets, unclear edges, and noisy images. Using work-in-progress automation software, experiments were conducted with models based on 2D SegNet, 2D and 3D U-Nets as preliminary research for a dental segmentation automation tool. The 2D U-Net with adjacent images demonstrates higher global accuracy of 0.82 than naïve U-Net variants. The 2D SegNet showed the second highest global accuracy of 0.96, and the 3D U-Net showed the best global accuracy of 0.99. The automated canal detection system through deep learning will contribute significantly to efficient treatment planning and to reducing patients' discomfort by a dentist. This study will be a preliminary report and an opportunity to explore the application of deep learning to other dental fields.

Figure 1

Preprocessing steps such that just the mandibular part remained.

Figure 2

Architecture of deep learning networks. (A) SegNet; (B) U-Net with fewer filters than the original U-Net (C) U-Net with the original number of filters; (D) 3D U-Net.

Figure 3

Training progress of each network. (A) 2D SegNet; (B) 2D U-Net; (C) 3D U-Net Each model was stopped when its training loss converged (A) Training loss of pre-trained SegNet with 600 epochs; (B) Training loss of non-pre-trained SegNet with 600 epochs; (C) Training loss of U-Net with 600 epochs. (D) Training loss of 3D U-Net with 100 epochs.

Figure 4

Segmentation result in the slice containing 2^nd molar. From left to right, test input image (A) ground truth mask; (B) 2D SegNet segmentation result; (C) 2D U-Net segmentation result; (D) 3D U-Net segmentation result.

Figure 5

Segmentation result in the slice containing mandibular foramen area. From left to right, test input image (A) ground truth mask; (B) 2D SegNet segmentation result; (C) 2D U-Net segmentation result; (D) 3D U-Net segmentation result.