IIUM Repository (IREP)

Supervised pterygium fibrovascular redness grading using generalized regression neural network

Che Azemin, Mohd Zulfaezal and Hilmi, Mohd. Radzi and Mohd. Kamal, Khairidzan (2014) Supervised pterygium fibrovascular redness grading using generalized regression neural network. In: 13th International Conference on New Trends in Intelligent Software Methodology Tools, and Techniques (SoMeT_14), 22-24 Sept. 2014, Langkawi, Malaysia.

[img] PDF - Published Version
Restricted to Repository staff only

Download (739kB) | Request a copy
PDF - Published Version
Download (448kB) | Preview


Pterygium is a growth on the eye that can cause blindness, with countries closer to the equator showing higher rate of incidence. However, there is a lack of research to study the severity and properties of the tissue. We propose the use of Generalized Neural Network (GRNN) to objectively quantify redness of the fibrovascular tissue. Comparative analysis using multiple feature selection algorithms indicates that error can be minimized when use with optimal set of features and suitable GRNN spread parameter. Features nominated by Minimum Redundancy Maximum Relevance gives the best performance with SSE = 3.55 and GRNN spread = 0.47.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Additional Information: 6768/40795 (New Trends in Software Methodologies, Tools and Techniques, H. Fujita et al. (Eds.), IOS Press, 2014. (ISBN: 978-1-61499-433-6, e-ISBN: 978-1-61499-434-3)
Uncontrolled Keywords: Image color analysis, generalized regression neural network, supervised learning.
Subjects: R Medicine > RE Ophthalmology
T Technology > TA Engineering (General). Civil engineering (General) > TA164 Bioengineering
Kulliyyahs/Centres/Divisions/Institutes: Kulliyyah of Allied Health Sciences > Department of Optometry and Visual Science
Kulliyyah of Medicine > Department of Ophthalmology
Depositing User: Dr. Mohd Zulfaezal Che Azemin
Date Deposited: 26 Jan 2015 11:31
Last Modified: 19 Jun 2018 09:41
URI: http://irep.iium.edu.my/id/eprint/40795

Actions (login required)

View Item View Item


Downloads per month over past year