Partial iris image recognition using wavelet based texture features

Now a day's security is the key issue all over in the world and iris biometric is shown to be the most authentic and accurate for human identification. In this paper, an effective partial iris image recognition is described based on wavelet texture features where capturing the entire iris may not be feasible. Daubechies 4-level Db8 wavelet transform is applied to two small portions namely left and right sub-images of the irises to extract distinctive iris features which is comprised of only 510 bits whereas existing approaches extracted as many as 1024 bits to represent the unique feature vector. Hamming distance similarity measurement technique was employed for recognition. The experimental results show that it is possible to use only significant portion (left and right sub-image) of the irises for human identification.