TY - GEN
T1 - Automatic Optic Disc Localization Using Particle Swarm Optimization Technique
AU - Subramanya Jois, S. P.
AU - Harsha, S.
AU - Harish Kumar, J. R.
N1 - Funding Information:
This work is supported by MHRD, Government of India, under the IMPRINT India Initiative.
Publisher Copyright:
© 2018 IEEE.
Copyright:
Copyright 2019 Elsevier B.V., All rights reserved.
PY - 2019/2/22
Y1 - 2019/2/22
N2 - There is a growing need for plenarily automated algorithms that expeditiously localize the optic disc region in retinal fundus images for the analysis of retinal pathologies such as glaucoma. In this paper, we propose a methodology based on particle swarm optimization for automatic localization of optic disc region from retinal fundus images, where minimization of the fitness function is utilized to resolve optimization quandaries. Here, kernels are modeled as particles and they test the region-of-interest based on the fitness function, in the respective databases, where it is likely that the optic disc exists. The proposed method is validated on a total of 1670 fundus images obtained from various publicly available fundus image datasets. The optic disc localization accuracy obtained by the proposed method are 100%, 98.01%, 96.15%, 98.87%, 100%, and 100% on DRIVE, DRISHTI-GS, DIARETDB0, DIARETDB1, DRIONS-DB, and MESSIDOR fundus image databases, respectively. The precision of localization was improved with initialization of kernel particles within bright region-of-interest in fundus images.
AB - There is a growing need for plenarily automated algorithms that expeditiously localize the optic disc region in retinal fundus images for the analysis of retinal pathologies such as glaucoma. In this paper, we propose a methodology based on particle swarm optimization for automatic localization of optic disc region from retinal fundus images, where minimization of the fitness function is utilized to resolve optimization quandaries. Here, kernels are modeled as particles and they test the region-of-interest based on the fitness function, in the respective databases, where it is likely that the optic disc exists. The proposed method is validated on a total of 1670 fundus images obtained from various publicly available fundus image datasets. The optic disc localization accuracy obtained by the proposed method are 100%, 98.01%, 96.15%, 98.87%, 100%, and 100% on DRIVE, DRISHTI-GS, DIARETDB0, DIARETDB1, DRIONS-DB, and MESSIDOR fundus image databases, respectively. The precision of localization was improved with initialization of kernel particles within bright region-of-interest in fundus images.
UR - http://www.scopus.com/inward/record.url?scp=85063237541&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85063237541&partnerID=8YFLogxK
U2 - 10.1109/TENCON.2018.8650053
DO - 10.1109/TENCON.2018.8650053
M3 - Conference contribution
AN - SCOPUS:85063237541
T3 - IEEE Region 10 Annual International Conference, Proceedings/TENCON
SP - 1718
EP - 1722
BT - Proceedings of TENCON 2018 - 2018 IEEE Region 10 Conference
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE Region 10 Conference, TENCON 2018
Y2 - 28 October 2018 through 31 October 2018
ER -