Research on the algorithms of iris recognition
Cuiping Xu · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2006
Iris recognition is a very reliable method for personal identity verification.To improve upon conventional algorithms related to iris recognition,methods are proposed for available region selection,texture feature extraction,and code matching of the iris.On the selection of available regions,a new division method was applied according to the biometric feature of the iris itself to obtain more information.In the area of texture feature extraction,the transforms of a Gabor wavelet are introduced.Dividing the frequencies of Gabor into two bands,different Gabor scale parameters were selected in every band,and the appropriate location parameters were chosen.Finally,in order to resolve the effects of iris image rotation on the overall result of iris recognition,it was found that the binary iris codes achieved must be compared using a method of shifting in a fixed length.Experimental results show that the proposed iris recognition methods perform better than conventional methods.