Tooth detection for each tooth type by application of faster R-CNNs to divided analysis areas of dental panoramic X-ray images.

This study aimed to propose a computerized method for detecting the tooth region for each tooth type as the initial stage in the development of a computer-aided diagnosis (CAD) scheme for dental panoramic X-ray images. Our database consists of 160 panoramic dental X-ray images obtained from 160 adult patients. To reduce false positives (FPs), the proposed method first extracts a rectangular area including all teeth from a dental panoramic X-ray image with a faster region using a convolutional neural network (Faster R-CNN). From the rectangular area including all teeth, six divided areas are then extracted with Faster R-CNN: top left, top center, top right, bottom left, bottom center, and bottom right. Faster R-CNNs for detecting tooth regions for each tooth type were trained individually for each of the divided areas that narrowed down the target tooth types. By applying these Faster R-CNNs to each divided area, the bounding boxes of each tooth were detected and classified into 32 tooth types. A k-fold cross-validation method with k = 4 was used for training and testing the proposed method. The detection rate for each tooth, number of FPs per image, mean intersection over union for each tooth, and classification accuracy for the 32 tooth types were 98.9%, 0.415, 0.748, and 91.7%, respectively, showing an improvement compared to the application of the Faster R-CNN once to the entire image (98.0%, 1.194, 0.736, and 88.8%).

Segmentation of teeth in panoramic dental X-ray images using U-Net with a loss function weighted on the tooth edge.

Panoramic dental X-ray imaging is an established method for the diagnosis of dental problems. However, the resolution of panoramic dental X-ray images is relatively low. Thus, early lesions are often overlooked. As the first step in the development of a computer-aided diagnosis scheme for panoramic dental X-ray images, we propose a computerized method for the segmentation of teeth using U-Net with a loss function weighted on the tooth edge. Our database consisted of 162 panoramic dental X-ray images. The training dataset consisted of 102 images, while the remaining 60 images were used as the test dataset. The loss function obtained by the cross entropy (CE) in the entire image is usually used in training U-Net. To improve the segmentation accuracy of the tooth edge, a loss function weighted on the tooth edge is proposed by adding the CE in the tooth edge region to the CE for the entire image. The mean Jaccard index and Dice index for U-Net with the loss function combining the CEs for the entire image and tooth edge were 0.864 and 0.927, respectively, which were significantly larger than those for U-Net with the CE for the entire image (0.802 and 0.890, p < 0.001) and U-Net with the CE for the tooth edge (0.826 and 0.905, p < 0.001). U-Net with the new loss function exhibited a higher segmentation accuracy of the tooth in panoramic dental X-ray images than that obtained by U-Net with the conventional loss function.