Iris recognition and classification methods

Abstrak

Conventional authentication methods such as passwords and Personal Identification Numbers (PINs) have proven to be vulnerable to misuse and data breaches, highlighting the urgent need for more reliable identity verification systems. Biometric-based authentication, particularly iris recognition, has emerged as a promising solution due to the unique and physiologically stable nature of iris patterns throughout a person's lifetime. However, previous studies in iris recognition have reported limited accuracy and lacked comprehensive evaluation using metrics such as Precision, Recall, and F1-Score. This study designed and implemented an iris recognition system using Gabor Wavelet as the feature extraction method combined with the K-Nearest Neighbor (K-NN) classification algorithm. The dataset was obtained from the CASIA Iris Database, consisting of 100 images covering 20 identity classes, of which 20 images were used as test data. The system pipeline comprised iris segmentation, normalization, feature extraction using Gabor Wavelet, and classification using K-NN with K=1 and Euclidean distance. System performance was evaluated using accuracy, Precision, Recall, and F1-Score metrics across multiple threshold values. The experimental results showed that the proposed system achieved an accuracy of 95%. At a threshold of 0.5, the system produced the best overall performance with a Precision of 0.90, Recall of 0.95, and F1-Score of 0.92. These findings confirmed that the combination of Gabor Wavelet and K-NN was effective for biometric-based iris recognition systems.