Densenet Model for Binary Glaucoma Classification Performance Assessment with Texture Feature

Authors

Wildan Jameel Hadi, Department Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq.Follow
Amal Sufiuh Ajrash, Department Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq.Follow
Sahar Muneam Salman, Department Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq.Follow
Mays jalal jasim, Department Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq.Follow
Mina Taha Ibrahim, Department Computer Science, College of Science for Women, University of Baghdad, Baghdad, Iraq.Follow

Abstract

The retina is an important portion of the eye because images of it are used by doctors to diagnose numerous eye diseases such as glaucoma, diabetic retinopathy, and cataracts. Indeed, segmented retinal imaging is a powerful tool for detecting unusual growths in the eye area as well as determining the size and structure of the optic disc. The combination of digital image processing and deep learning techniques enables the development of automated approaches for detecting glaucoma. Within this framework, the objective of this study is to achieve prompt detection of glaucoma by utilizing a DenseNet-121 Network in conjunction with texture qualities derived from local binary patterns (LBP). The proposed method can be categorized into four steps: (i) obtaining images from the OIH public database; (ii) preprocessing the images by extracting texture attributes using LBP; (iii) classifying glaucoma images and normal images using a DenseNet-121 network; and (iv) validating the proposal based on performance metrics. Based on the results of the proposed strategy, the accuracy remains approximately 96%.

Keywords

Binary Classification, Dense Net, Eye Diseases, LBP, Texture.

Article Type

Article

How to Cite this Article

Jameel Hadi, Wildan; Sufiuh Ajrash, Amal; Muneam Salman, Sahar; jalal jasim, Mays; and Taha Ibrahim, Mina (2024) "Densenet Model for Binary Glaucoma Classification Performance Assessment with Texture Feature," Baghdad Science Journal: Vol. 21: Iss. 12, Article 18.
DOI: https://doi.org/10.21123/bsj.2024.9857