ENHANCING RETINAL DISEASE SCREENING WITH LOCAL BINARY PATTERNS

Abstract

When sugar level (glucose) in the 
blood fails to regulate the insulin properly 
in human body, diabetic is occurred. The 
effect of diabetic on eye causes diabetic 
retinopathy. Diabetic Retinopathy is one of 
the complicated diabetes which can cause 
blindness. It   is   metabolic   and   
the 
disordered   
patients  
perceive    
no 
symptoms until the disease is at late stage. 
So early detection and proper treatment 
has to be ensured.  To  serve this 
purpose, various automated systems have 
been designed).  A  key   feature    to 
recognize Diabetic Retinopathy is to 
eye.   
detect Microaneurysm in the fundus    of 
the  
This    
work   investigates 
discrimination capabilities in the texture of 
fundus images to differentiate between 
pathological and healthy images. For this 
purpose, the performance of Local Binary 
Patterns (LBP) as a texture descriptor for 
retinal images has   
Transform to improve the red lesion 
detection. Finally the output image was 
classified 
as 
retinopathy (DR). 
Normal and Diabetic 
These results suggest that the method 
presented in this paper is a robust algorithm 
for describing retina texture and can be 
useful in a diagnosis aid system for 
retinal disease screening.