Iris Recognition Based on Pulsed Coupled Neural Networks

Abstract

Based on Pulsed Coupled Neural Network Model, respectively to study the one-dimensional time series, row time series, column time series will keep some invariance by image migration. To apply this nature in iris recognition field, found that to extraction features using row time series method, which can resist rotation changes between the same samples, and keep enough details to distinguish between different samples. Compared with the traditional iris recognition method which using iris phase structure encoded, the results show that row time series method, is less sensitive for phase, and can resist rotation changes better, is effective.