10-06-2017, 06:56 AM
An improved new approach to identifying the person using the iris recognition technique. This approach is based on the Artificial Neural Network and Vector Support Machine (SVM) as an iris pattern classifier. Prior to the classifier, the region of interest, ie the iris region, is segmented using the Canny edge detector and the Hough transform. Provided that the effect of the eyelid and the eyelashes is reduced. The Daugman rubber sheet model is used to obtain the standardised iris to improve computational efficiency and proper dimensionality. In addition, the discriminant feature sequence is obtained by extracting segmented iris image features using a 1D Log Gabor wavelet. The coding is performed using the phase quantification to obtain characteristic vectors. These binary sequence function vectors are used to form SVM and ANN as iris pattern classifier. Experimental tests are performed on the basis of CASIAIrisV4 standard data.