IRIS RECOGNITION SYSTEM FOR SCALE AND ROTATION INVARIANCE WITH SVM CLASSIFIER
Ashna Mary George, C. Anand, Deva Durai · 2013
Abstract— Biometric technology is used for the automatic recognition of an individual based on the unique characteristics or features possessed by an individual. Many Biometric Technologies are available today. Iris Recognition System is the most consistent and truthful Biometric identification system available. Iris Recognition is the recognition of an individual based on iris features. In the past few years many methods are used to improve the performance of iris recognition systems. These methods mainly focused on the robustness, accuracy and rapidity of iris recognition systems. In this paper some new techniques are used to improve the overall performance of iris recognition systems. First, this paper proposes a new technique for segmentation by segmenting three subregions of ROI for eliminating the noises. Second, after 2-D Gabor filtering multiscale and multiorientation feature extraction is used to extract features from each subregion. The extracted features from each subregion are fused using simple sum fusion. Finally, SVM classifier is used for recognition which accelerates huge search in iris databases. The performance was evaluated on popular iris database and the experiment results shows that these techniques are experimentally more robust and accurate with less computation time compared with most existing techniques.