Texture Feature Analysis of Digital Fundus Images for Early Detection of Diabetic Retinopathy

Author

Ashraf, Muhammad Nadeem ; Habib, Zulfiqar ; Hussain, Mutawarra

Author_Institution

Dept. of Comput. Sci., COMSATS Inst. of Inf. Technol., Lahore, Pakistan

fYear

2014

fDate

6-8 Aug. 2014

Firstpage

57

Lastpage

62

Abstract

Diabetic retinopathy (DR) is a complication where the retina of a diabetic patient is damaged due to fluid leakage from the blood vessels into the retina and the patient may suffer from complete blindness if untreated. Hemorrhages and Microaneurysms (HMAs) are the early signs that appear in retina at the initial stage of DR. Early diagnosis of HMAs is crucial to prevent blindness and fundus image is used for this purpose. We have focused on the analysis of texture micro-patterns of the regions of interest (ROIs), which are suspicious regions in a fundus image, for the detection of HMAs. Texture micro-structures of ROIs are analyzed through Local Binary Pattern (LBP) for their description. Finally Support Vector Machine (SVM) is employed to identify whether an ROI contains HMAs or not.

Keywords

blood vessels; diseases; eye; haemorheology; image texture; medical image processing; patient diagnosis; support vector machines; HMA; LBP; ROI; SVM; blindness; blood vessels; diabetic patient; diabetic retinopathy; digital fundus images; early detection; fluid leakage; hemorrhages; local binary pattern; microaneurysms; regions of interest; support vector machine; texture feature analysis; texture micro-patterns; texture microstructure; Accuracy; Diabetes; Feature extraction; Retina; Retinopathy; Support vector machine classification; Diabetic retinopathy (DR); Fundus image; Local binary pattern (LBP); Support vector machine (SVM);

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Graphics, Imaging and Visualization (CGIV), 2014 11th International Conference on

Conference_Location

Singapore

Type

conf

DOI

10.1109/CGiV.2014.29

Filename

6934121