Classification of Gender Individual Identification Using Local Binary Pattern on Palatine Rugae Image

Abstract

Major disasters caused many casualties with the condition of the damaged bodies. It causes the individual identification process to be ineffective through biometric characteristics (such as lips and fingerprints). However, the palatine rugae can carry the individual identification process. Palatine rugae have unique and individual characteristics and are more resistant to trauma because of their internal location. In this study, an individual identification system is proposed to identify gender using the image of palatine rugae. The proposed system is developed by several algorithms and methods, such as Local Binary Pattern (LBP) as the feature extraction method and K-Nearest Neighbor (KNN) as the classification method. Based on the result of the system performed test, the proposed system can identify the gender of an individual by the combination of recognized palatine rugae patterns. The system achieved an accuracy test result of 100% with a specific configuration of LBP and KNN. The research contribution in this study is to develop the individual gender identification system, which proceeds with the palatine rugae pattern image with unique biometric characteristics as an input. The system applied several methods and algorithms, such as Geometric Active Contour (GAC) as a segmentation algorithm, Local Binary Pattern (LBP) as a feature extraction method, and K Nearest Neighbor (KNN) as a classification method.