An iris recognition method based on 2DWPCA and neural network
Zhou Zhiping, Maomao Hui, Ziwen Sun · 2009
An iris recognition method based on two-dimensional weighted principal component analysis (2DWPCA) and adaptive artificial neural network is proposed. As different iris region contains different recognition information, different weighting value is allocated to different region after compensating illumination intensity of the image in preprocessing. The two-dimensional principal component analysis is used to calculate the weighted subspace. And then 2DWPCA is utilized to extract the feature. Adaptive artificial neural network is employed to train and recognize the generated feature vectors. Owing to the 2DPCA features optimization of 2DPCA extraction and the self-adaption of neural network, the recognition ratio and robustness were greatly improved.