Abstract :
Iris recognition is a very reliable method for personal identity verification. Some pre-processing methods are first discussed in this paper, such as iris localization, iris image enhancement etc. To improve some lacks in the published algorithms of iris recognition, some effective methods are proposed for the available region selection, texture feature extraction and code matching of iris. On the selection of available regions, a new division method is applied according to the biometric feature of iris itself to obtain more information. On the texture feature extraction, the transforms of the Gabor wavelet is introduced. Dividing the frequencies of Gabor into two bands, different Gabor scale parameters are selected in every band, and the appropriate location parameters are chosen. In order to resolve the effects of iris image rotation on the result of iris recognition, the binary iris codes achieved must be compared using the method of shifting in a fixed length. Experimental results show that the iris recognition method proposed has a better performance
Keywords :
feature extraction; image recognition; image texture; wavelet transforms; Gabor wavelet transforms; code matching; iris image enhancement; iris image rotation; iris localization; iris recognition algorithms; personal identity verification; texture feature extraction; Automation; Biometrics; Eyelashes; Eyelids; Feature extraction; Gabor filters; Image resolution; Iris recognition; Nonlinear filters; Wavelet transforms; Gabor Wavelet; Iris texture; feature extraction; rotation;
