Feature identification and classification of hand based biometrics through ensemble learning approach

In the last decade identification using biometric system has been growing rapidly. This research shows the new technique based on biometric recognition using hand structure. A database with 200 sample of human having both sides of hand is used. The technique used according to the human comfort while taking pictures and gives segmentation of fingers and hands with high efficiency. Almost, 94 features have been extracted from the given datasets and various classification and techniques have been assessed. The k-cross validation rule has been used to avoid the problem of overfitting and underfitting and best sampling method is also used. The experimented results show the better accuracy of our system compared to other existing systems. The introduced system is also able to show good results using generic features as feature extraction method and SVM as classification method. General terms: Classification, Pattern Recognition, Security, Algorithms. © 2022 The Authors