Machine Learning based diagnosis of Diabetic Retinopathy using digital Fundus images with CLAHE along FPGA Methodology

  • Yallanti Sowjanya Kumari, Mekala Srinivasa Rao, Ranga Swamy Sirisati

Abstract

In a diabetic patient when small nervous of retina are being damaged which are observed between posterior part of eye is termed as Diabetic Retinopathy (DR). In aged population who are working Diabetic retinopathy is considered as the main cause of blindness. Though treatments are available to some extent the detection of it is failed some times. If it is detected at an early stage then it can be treated well to get good results in diabetic patients. Beside this early detection is also helpful in order to slow the disease progression by controlling the risk factors which are modifiable such as blood pressure, blood glucose etc. In DR PDR and NPDR are two main stages. In order to verdict and carry out the treatment of eye diseases digital retinal fundus images play a vital role. With the help of biomicroscopy by senior ophthalmologists diabetic retinopathy can be detected well. In proposed method we have used the combination of CLAHE along FPGA in order to get a high resolution images at last which are helpful in categorizing the exact stage of diabetic retinopathy with detection of exact areas contrast, hard exudates and area of the blood vessels. Beside this with proper treatment is carried out. In proposed method a dataset of digital fundus images are considered and with the help of required classifiers in machine learning the exact stage of the disease is recognized. A critical comparison of various classifiers is carried out in order to observe the obtained high accuracy. The obtained high accuracy results along with the high quality images are then considered in order to predict the exact stage of diabetic retinopathy to take necessary and correct measure either to constrain the disease or to carry out necessary operation.

 

Keywords: Adaptive histogram equalization (AHE), Contrast Limited Adaptive Histogram Equalization (CLAHE), Deep Learning (DL), Diabetic Retinopathy (DR), (Field Programmable Gate Arrays) FPGA, Intra Retinal  Microvascular Abnormalities (IRMA), Proliferate Diabetes Retinopathy (PDR), Machine Learning (ML), Non- Proliferate Diabetic Retinopathy (NPDR) , SVM (Support Vector Machine).

Published
2020-06-06
How to Cite
Yallanti Sowjanya Kumari, Mekala Srinivasa Rao, Ranga Swamy Sirisati. (2020). Machine Learning based diagnosis of Diabetic Retinopathy using digital Fundus images with CLAHE along FPGA Methodology. International Journal of Advanced Science and Technology, 29(05), 12748-12759. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/25875