Palmprint Identification Using Dolphin Optimization

Abstract

Palm print recognition is a rapidly evolving area in the field of biometrics, providing a high level of security for various applications Advances in scanning technology and software have led to faster and more accurate palm print analysis, we have proposed in this paper a system for palm recognition using the Dolphin Optimization Algorithm (DOA) as a computational technique inspired by nature aimed at solving complex improvement problems. We reduced the number of image features using the Histograms of oriented gradients (HOG) algorithm, we named this method as (DOA) and we also proposed a hybrid method by integrating the DOA algorithm with the Support Vector Machine (SVM) model to improve prediction accuracy by combining DSA's ability to search for global optimal solutions with the effective classification capabilities of SVM, this allows The hybrid approach creates a robust and the proposed hybrid method was named (SVM-DOA), and we also proposed a hybrid method by integrating the DOA algorithm with fuzzy c – mean (FCM) and we named the proposed hybrid method (DOA-fuzzy membership), we verified the validity of the proposed method on public database images of palm print. experiment show that the average accuracy rate of the dolphin swarm algorithm (DSO) is (96.8%), while the average accuracy rate of the proposed hybrid algorithm (DSO-SVM) is (97.8%), and the average accuracy of the proposed hybrid algorithm (DSO-fuzzy membership) is (98.1%).