|
Prevention of Non-Proliferative Diabetic Retinopathy (NPDR) by Detecting Exudates, Sub-conjunctival Haemorrhages and Micro-aneurysms
|
|
| Author:
|
K. MANJU, KOLLA SRUTHI, P. POOJA KARAN, M. POOJA, M. AMUTHA
|
| Abstract:
|
Screening and delivering treatment to the ophthalmic disease like Diabetic Retinopathy is a tedious and highly risky process because the longer one has diabetes, the greater is the risk of developing diabetic retinopathy. Also, the need for a physician to meet the requirement is impossible. So it has been proposed a method that adds a solution to the problem by readily diagnosing the patient to detect NPDR (Non – Proliferative Diabetic Retinopathy) which is a prior stage of DR so that loss of vision can be prevented. The experimental analysis deals with the study fundus images which includes both positive and negative NPDR as input image which undergoes vessel segmentation in MATLAB platform is primarily finding Exudates, Haemorrhage and Micro-aneurysms. After morphological operations like pre- processing steps, segmenting, masking, arithmetical operations on the fundus image, the exudates and haemorrhages can be detected. Based on the study on a different set of DRIVE (Digital Retinal Image for vessel Extraction) images, the resultant processed image can be interpreted with the visible information about the complexity of the ailment which also includes the healthy samples. The parametric analysis of Exudates, Haemor- rhages and Aneurysms concerning the threshold values through which NPDR can be predicted. The output of the proposed method describes whether the person subjected to test have the abnormalities for NPDR or not. This also interpreted by the study results on various DRIVE datasets to arrive at this accuracy.
|
| Keyword:
|
Non-Proliferative Diabetic Retinopathy, Diabetes, Exudates, Haemorrhage, Micro-aneurysms, Morphological Operations and DRIVE Images.
|
| EOI:
|
-
|
| DOI:
|
https://doi.org/10.31838/ijpr/2020.12.01.222
|
| Download:
|
Request For Article
|
|
|