Abstract
Tooth enamel is the hardest tissue in the human body, formed by prism layers in regularly alternating directions. These prisms form the Hunter-Schreger Band (HSB) pattern when underside illumination, which is composed of light and dark stripes resembling fingerprints. Hunter-Schreger Band patterns are highly variable and unique for each tooth and can be used for personal identification. Hunter-Schreger Band images of 115 teeth were captured with a Canon EOS 5D mark III camera coupled to an InfiniProbe TS-160 lens (Infinity Photo-Optical Company, Boulder, Co, USA). The algorithms were developed in Python 3.9 using image-specific libraries NumPy, Scikit-Image, OpenCV, SciPy and TensorFlow. The evaluation of performance of HSB filtering and binarization was performed visually and indirectly. The feature extraction and matching algorithms were created upon adaptations from fingerprint techniques and revealed as the result of matching evaluation an equal error rate of 0.061, when both enlightened sides of each tooth were used together in a sample of 115 extracted teeth. The pipeline developed here allows the capture, image treatment, and matching among the images stored in the database. This demonstrated the potential of the new biometric trait.