Object detection in dental X-ray image using 5-axis coordinate system

X-rayimages are critical data sources analyzing the conditions of the teeth, gums, jaws, and bone structure of the mouth. Teeth recognition is one of the most fundamental studies in image processing-based diagnosis. Most existing recognition studies only consider the 4-axis of an object detection model because they address normal object detection while standing on the ground. However, considering the shape of teeth with various orientations, the existing 4-axis based model leads to inaccurate and inefficient recognition results. Thus, in this study, we propose a 5-axis based object detection model that considers the orientation of the teeth. Based on the teeth image dataset labeled with 5-axis ground truth, our proposed method processed 5-axis annotated data with a variant of the region-based convolutional neural network (R-CNN architecture). In the experiment, our proposed method outperformed the existing 4-axis approach in qualitative and quantitative results. Based on the experimental results, we conclude that the 5-axis based recognition model proposed in this study will be an important basis for dental image-based diagnosis.