Backpropagation Design for Authenticating Blood Vessel Patterns of the Back of the Hand Using GLRLM

Abstract

Digital security is a critical aspect in the current era of information technology, where access to personal devices and data is often the main target by irresponsible parties. Traditional identification methods such as passwords and PINs are starting to show limitations in addressing increasingly complex security challenges.. The dorsal hand veins offer certain advantages that make them an attractive option for biometric recognition systems because the dorsal hand vein pattern tends to be stable over time, unaffected by external factors such as changes in weather or hygiene. This research aims to develop a system that can identify the blood vessels of the back of the hand as a biometric sign. The approach used involves extracting GLRLM features and applying the Back Propagation Neural Network identification method. The main goal is to achieve a higher level of accuracy than previous studies in the same domain. The identification process involves several stages, starting from image reception, image pre-processing, segmentation, feature extraction, identification, to obtaining images resulting from blood vessel identification. Test results show that the system developed achieved an average success rate of 82.52% based on five different test scenarios. The fourth scenario was proven to provide the highest test accuracy results, namely 87%.