The iris segmentation step is usually the most time consuming stage of biometric systems when dealing with non ideal conditions, which produce diverse noise factors during the acquisition. On the other side, it also represents a crucial step since poor removal of noise leads to degradation of recognition performance. In this work, a lightweight fuzzy-based solution has been explored. The goal is to propose a fast but reliable segmentation approach which preserves the original resolution of the iris images. The preliminary results obtained on a subset of MICHE dataset, confirmed both acceptable performance in terms of time consumption and good quality of segmentation mask suitable for matching purposes.
A Lightweight Mamdani Fuzzy Controller for Noise Removal on Iris Images
Abate, Andrea Francesco;Nappi, Michele;Narducci, Fabio
2017-01-01
Abstract
The iris segmentation step is usually the most time consuming stage of biometric systems when dealing with non ideal conditions, which produce diverse noise factors during the acquisition. On the other side, it also represents a crucial step since poor removal of noise leads to degradation of recognition performance. In this work, a lightweight fuzzy-based solution has been explored. The goal is to propose a fast but reliable segmentation approach which preserves the original resolution of the iris images. The preliminary results obtained on a subset of MICHE dataset, confirmed both acceptable performance in terms of time consumption and good quality of segmentation mask suitable for matching purposes.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
