Iris recognition in visible spectrum based on multi-layer analogous convolution and collaborative representation

摘要

With the development of mobile Internet and intelligent mobile devices, mobile iris recognition is a new advanced frontier for identification. This paper presents an iris recognition algorithm based on a multi-layer analogous convolutional structure and collaborative representation to solve the high intra-class difference caused by the visible lighting interference and the change of image acquisition sensors. The proposed method uses the multi-layer analogous convolution to reduce the dimension of iris texture information with lower computational complexity. Then, the feature of the obtained texture map is extracted and classified by collaborative representation method. The experimental analysis on MICHE-I demonstrates the adaptability of proposed method to changes in lighting condition and in acquisition sensors. The experimental results show that our algorithm provides better accuracy, AUC and EER compared to the iris recognition method such as Gabor, PCA, SRC, MACS + SRC and CRC. (C) 2018 Elsevier B.V. All rights reserved.