Available online www.ejaet.com European Journal of Advances in Engineering and Technology, 2015, 2(1): 55-64 Research Article ISSN: 2394 - 658X 55 Teeth Feature Extraction and Matching for Human Identification Using Scale Invariant Feature Transform Algorithm Dipali Rindhe and Ganesh Sable Savitribai Phule Women Engineering College, Aurangabad, Maharashtra, India dipalirindhe8943@gmail.com _____________________________________________________________________________________________ ABSTRACT Dental biometrics has emerged as vital biometric information of human being due to its stability, invariant nature and uniqueness. Dental radiograph and dental photograph are tools mostly used in biometrics as it provides information about teeth in detail. The proposed system has four main stages as preprocessing, feature extraction, feature matching and finalized recognized person. Preprocessing stage uses conversion and segmentation of quiery input image to make it fesible for feature extraction. Feature extraction uses SIFT algoritham as it extract highly distinctive invariant features which uses for matching purpose. In matching input image is comparatively matched with every image of database and find out the maximum matched features image of person. The system is work for both types of dental images i.e. photograph and radiograph in which two different datasets are required. The required database contains 50 images of dental photographs and 50 images of dental radiographs so experimentation has done on total 100 images and that are taken from Dyanita dental clinic. The proposed system is implemented in Matlab/ R2012a programming tool. Key words: Dental biometrics, dental photograph, dental radiograph _________________________________________________________________________________ INTRODUCTION Teeth are parts of human organ that are not easily decayed and located inside mouth. It has its own characteristics based on a number of distinctive features for each individual tooth. Therefore, teeth based identification is one of reliable tools for human identification. In general, human has 32 teeth and each tooth has five surfaces it means that inside a mouth there are 160 tooth surfaces with various conditions [4]. If we use dental features as a tool of identification, manual matching based on teeth appearance needs a large amount of time and some expertise, therefore computer aided for an identification system is needed. The features of teeth includes properties of the teeth e.g. shape and size of teeth, crown and root morphology, pathology, and dental restorations, periodontal tissue features and anatomical features. During the feature extraction certain salient information of teeth such as contour, artificial prosthesis, shape and size of teeth number of cuspids etc is extracted from dental image i.e. dental photograph and radiograph [9]. In this work the feature extracted is tooth contour and tooth shape because they remains more invariant over time as compared to some other features of teeth and this thing play an important role in dental biometric. Tooth Anatomy A tooth (plural teeth) is a small, calcified, whitish structure found in the jaws (or mouths) of many vertebrates and used to break down food. The roots of teeth are covered by gums. Teeth are not made of bone, but rather of multiple tissues of varying density and hardness. A typical healthy tooth has number of layers and its internal structure. Each tooth relates to the gum and surrounding jaw bone. Fig. 1 shows the tooth anatomy. The crown is the part of the tooth that is visible above the gum (gingiva) and neck is the region of the tooth that is at the gum line, between the root and the crown. The root is the region of the tooth that is below the gum. Some teeth have only one roots, for example, incisors and canine (‘eye’) teeth, whereas molars and premolars have 4 roots per tooth.