Algoritmo para Pupilometria Dinâmica Baseado em Crescimento de Regiões

摘要

The difficulty of performing secure personal au-thentication has lead for the creation of automatic identification systems based on biometric features. Methods of biometric identification by iris texture are quite efficient and don 't require direct contact of the person with an equipment. Furthermore, they are not invasive and have low error rate [1]. However, these tech-niques are susceptible to fraud because of the use of static im-ages. The dynamic pupillometry, joint with iris recognition, can make Identification even more secure by analyzing human ref-lexes. This article presents existing methods of iris segmentation for dynamic pupillometry, proposes a new approach and compare the obtained results.