Please use this identifier to cite or link to this item: https://hdl.handle.net/11147/14254
Title: An iris segmentation scheme based on bendlets
Authors: Aghazadeh, Nasser
Abbasi, Mandana
Noras, Parisa
Keywords: Bendlets
Image denoising
Iris detection
Iris segmentation
Image
Localization
Recognition
Publisher: Springer London Ltd
Abstract: Due to the effect of agents such as ambiance, transition channel, and other agents, images are polluted by noise during collection, transition, and compaction, leading to decrease image quality. Noise can decrease the accuracy of the next stages of image processing systems. Therefore, one of the vital stages in the novel processing systems is denoising. This article offers a novel image denoising approach using bendlets. Other multi-scale transformations (such as wavelets, curvelets, and shearlets) cannot recognize properties such as location, direction, and curvature of discontinuities well in piecewise stable images. To solve this problem, bendlets are suggested in this article. Bendlets differ from other multi-scale transformations in that an additional bending parameter is utilized for recognizing the curvature of discontinuities. Bendlets need a fewer number of coefficients to identify curvatures than other multi-scale transformations. Furthermore, they help to make the edges more obvious. The suggested approach is utilized on the UBIRIS.V2 database. It earns better accuracy and stability than other multi-scale transformations.
URI: https://doi.org/10.1007/s11760-023-02940-1
https://hdl.handle.net/11147/14254
ISSN: 1863-1703
1863-1711
Appears in Collections:Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
WoS İndeksli Yayınlar Koleksiyonu / WoS Indexed Publications Collection

Show full item record



CORE Recommender

Page view(s)

60
checked on Jun 17, 2024

Google ScholarTM

Check




Altmetric


Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.