Wrist hand vein recognition using local line binary pattern (LLBP)

The biometrics recognition system is now widely applied in various fields such as security systems and employee attendance. However, there are some weaknesses that the recognition system has, so that new biometrics recognition objects are developed, namely recognition based on the wrist hand vein image. Some feature extraction methods that have been used in vascular-based recognition systems such as Dense-LBP, RBF Network, Multi-Layer Perceptron (MLP) and Support Vector Machine (SVM) and Hessian-based result in poor accuracy, so that this study tries to apply the Local Line Binary Pattern (LLBP) method in the feature extraction stage because based on previous research, the Local Line Binary Pattern (LLBP) method is able to process the image at the stage feature extraction very well. The results of individual recognition experiments using the LLBP method were carried out using 50 individual data from datasets taken directly using infrared cameras, resulting an accuracy of 96.50. Accuracy obtained is influenced by the quality of the dataset used and in this study, it shows that the proposed LLBP method has good capability in individual recognition systems.