Raspberry Pi-Based Dual Iris Recognition System Using Coiflet Wavelet

This study developed a dual iris recognition biometric system using the 2-D Coiflet wavelet decomposition. The system mainly consists of the Raspberry Pi 4 Model B, Raspberry Pi Camera v2 NoIR, and NIR technology. The left and right irises of the subject are captured to be processed through segmentation, normalization, and iris feature extraction using PyWavelets. With the help of the CASIA-IrisV4 database, a multi-class support vector machine (SVM) model is trained to create a system that determines genuine and impostor iris images. After a total of seventy (70) attempts, with a False Accept Rate (FAR) of 0% and False Reject Rate (FRR) of 5%, the system yielded an accuracy of 95% accuracy. The usage of Coiflet wavelet has been deemed effective for iris recognition.