Numbering of Teeth in the Mixed Dentition Stage Using Deep Learning

Panoramic dental X-ray images are medical images that show the dentition status of the teeth in the mouth, dental diseases, and tooth deficiencies. Orthodontic treatment processes are shaped according to the results obtained visually from these images. Various systems that make inferences from these images have been developed to minimize errors in treatment processes. Within the scope of this study, a one-step deep learning approach is proposed to detect primary teeth and permanent teeth through panoramic dental X-rays and number the teeth according to the Universal Numbering System. In the study, determinations are made by implementing an existing deep learning model specifically for this problem. The numbering of the detected teeth is carried out with high performance by the deep model, and the need to use additional algorithms for tooth numbering is eliminated. The proposed approach detects and numbers the teeth on panoramic dental X-rays with 92% accuracy.