Ronaldo Martins da Costa,  Eveline B. Rodrigues,  Adilson Gonzaga. 

This work presents an efficient method for human iris segmentation using Euclidean DistanceTransform. The goal is providing data within a Region of Interest (ROI), eliminating eyelid and eyelash, for applications that recognize people through the iris. The method was considered efficient, because its success was about 80% of iris database. The method was considered fast, because its medium processing time is about 1 second per image, executing on MatLab. We have used the CASIA iris database provided by the National Laboratory of Pattern Recognition (NLPR), Institute of Automation (IA) and Chinese Academy of Sciences (CAS), with 756 iris images.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2005/0018.pdf

Processando consulta.