An Enhanced Approach for Iris Recognition Using Fusion of FWT with Gabor Wavelet Transform and Daugman Encoding

This paper discusses iris recognition, which is accepted as one of the best biometric methods for identifying an individual. A comparative analysis is done for eight algorithms, namely DCWT, FWT, SWT, CWT, DB, Complex dual tree, Haar wavelet and Wavelet packet for extracting the feature from iris image. Extracting the iris features of the image is still inconceivable as the existing algorithms fail to deliver the maximum accuracy. Among the eight algorithms, Fast Wavelet Transform (FWT) delivers 72.9 percent accuracy. This paper suggests a better way to enhance the accuracy using the Fast Wavelet Transform with the use of Gabor Wavelet Transform and Daugman algorithm. Gabor Wavelet Transform has multi-resolution and multi-orientation properties, which makes it popular for feature extraction. Daugman algorithm has computational simplicity and speed, which supposes the proposed method. Learning Vector Quantization (LVQ) neural network is used in the authentication unit. Results from confusion matrix and ROC shows that the proposed method produces cent percent accuracy. This implies that the effectiveness in authenticating the right person will be higher.