ISSN 1054-6618, Pattern Recognition and Image Analysis, 2020, Vol. 30, No. 1, pp. 125–133. © Pleiades Publishing, Ltd., 2020. Numbering and Classification of Panoramic Dental Images Using 6-Layer Convolutional Neural Network Prerna Singh a, * and Priti Sehgal b, ** a Department of Computer Science, University of Delhi, Delhi, 110007 India b Department of Computer Science, Keshav Mahavidyalaya, University of Delhi, Delhi, 110034 India * e-mail: prerna.singh@jimsindia.org ** e-mail: psehgal@keshav.du.ac.in Abstract—Deep Convolution Neural Network is one of the most powerful tools to solve complex problems of image classification, image recognition, financial analysis, medical diagnosis and many similar problems. A dental panoramic image consists of collection of teeth of both upper jaw and lower jaw. Automatic classifica- tion of dental panoramic images into various tooth types such as canines, incisors, premolars and molars has been a challenging task and involves crucial role of an experienced dentist. In this paper, we propose a tech- nique for numbering and classification of the panoramic dental images. The proposed algorithm consists of four stages namely pre-processing, segmentation, numbering and classification. The pre-processed pan- oramic dental images are segmented using fuzzy c-mean clustering and subjected to vertical integral projec- tion to extract a single tooth. The image dataset consists of 400 dental panoramic images collected from var- ious dental clinics. The 400 dental images are divided into 240 training samples and 160 testing samples. The image data set is augmented by applying various transformations. Panoramic dental images are further num- bered using a universal dental numbering system. Finally, the classification is done with the help of 6-layer deep convolution neural network (DCNN) consisting of 3 convolutional neural network and 3 fully con- nected network. The tooth is classified as canine, incisor, molar and premolar. An accuracy of 95% has been achieved for augmented database and 92% for original dataset with the proposed algorithm. The proposed numbering and classification of dental panoramic images is useful in biomedical application and postmortem recording of dental records. In case of big calamity, the system can also assist the dentist in recording post mortem dental record that is a very lengthy and arduous task. Keywords: deep convolution neural network, segmentation, tooth classification DOI: 10.1134/S1054661820010149 INTRODUCTION Teeth have a significant role in esthetics, mastica- tion and phonetics [1]. Automated dental identifica- tion system (ADIS) has been developed to extract information from the tooth images [2]. The technique has been distinctively used for identification of unknown persons by mapping the postmortem dental record with antemortem dental record. The system leads to saving time and gives accurate result. How- ever, the system may not be successful due to the long gap between antemortem and postmortem matching and changes in dental features. In-order to overcome these shortcomings, several works have been proposed for automation of dental tooth information from den- tal X-ray images for better matching between ante- mortem images and postmortem images [3–7]. In this paper we have used Deep Convolutional Neural Network for classification of dental images into molar, premolar, canine and incisor. According to the universal dental numbering system as shown in Fig. 2, a panoramic image consists of 32 teeth where 16 teeth are on the upper part of jaw and other 16 teeth are on the lower part of the jaw. The lower jaw is called mandible and the upper jaw is called maxilla [8]. Fig- ure 1 portrays the panoramic dental X-ray image and Figure 2 shows the universal dental numbering system. A panoramic dental image consists of four important components namely teeth, gums, pulp, and back- ground. The teeth are separated accordingly and the region of interest (ROIs) identified. Every tooth has crown and root, which constitute the upper and lower part of the tooth respectively. The use of deep convolutional neural network algorithms is helpful in organizing, analyzing, and interpreting the various images [9]. CNN is comprised of several convolutional layers followed by fully con- nected layers in order to effectively segment the image. CNN has three distinct layers namely convolution layer, pooling layer, and fully connected layer. CNN provides an extremely good performance in analysis of medical images. Deep learning methods when employed in CNN have helped in understanding the APPLIED PROBLEMS Received June 1, 2019; revised October 1, 2019; accepted October 21, 2019