Extraction of iris features using grey level co-occurrence matrix method with improved FAR and FRR for iris authentication in comparison with principal component analysis

• Main Authors: Jayanthi, Naga Guru Rahul, Nanmaran, R.
• Format: Conference Proceeding
• Language: English
• Subjects: Artificial neural networks; Feature extraction; Matrix methods; Performance enhancement; Principal components analysis; Rejection rate; Statistical analysis
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0150630

The accuracy of the iris authentication system can be improved by including more features from Iris videos or images. This research aimed at developing Iris authentication systems using different feature extraction methods and ANN classifiers. In this research a grey-level co-occurrence matrix(GLCM) based feature extraction technique is proposed to improve the performance of the iris authentication system. The proposed work is compared with another feature extraction technique called Principal component analysis(PCA) and the sample size for each group is 30. The Artificial neural network (ANN) is used for iris detection and the performance of iris authentication is measured using two parameters namely False acceptance rate (FAR) and False rejection rate (FRR). Grey-level co-occurrence matrix(GLCM) provides mean FAR (0.1743) and FRR (86.86) and PCA provides mean FAR (0.0154)and FRR (76.33). Based on the experimental results and statistical analysis using independent sample T test, the Grey-level co-occurrence matrix(GLCM) method based Iris authentication significantly performs better than Principal component analysis(PCA) method with FAR(P=0.005) and FRR(P=0.001).