Classifying Diabetic Retinopathy using Deep Learning Architecture

Abstract

A recent development in the state-of-art technology machine learning plays a vital role in the image processing applications such as biomedical, satellite image processing, Artificial Intelligence such as object identification and recognition and so on. In Global, diabetic retinopathy suffered patients growing vastly. And the fact is earliest stage could not diagnoses to normal eye vision. Increasing necessity of finding a diabetic retinopathy as earliest would stops vision loss for prolonged diabetes patient although suffered youngs'. Severity of the diabetic retinopathy disease is based on a presence of microaneurysms, exudates, neovascularization, Haemorrhages. Experts are categorized those diabetic retinopathy in to five stages such as normal, mild, moderate, severe Nonproliferative(NPDR) or Proliferative diabetic retinopathy patient(PDR). A proposed deep learning approach such as Deep Convolutional Neural Network(DCNN) gives high accuracy in classification of these diseases through spatial analysis. A DCNN is more complex architecture inferred more from human visual perspects. Amongst other supervised algorithms involved, proposed solution is to find a better and optimized way to classifying the fundus image with little pre-processing techniques. Our proposed architecture deployed with dropout layer techniques yields around 94-96 percent accuracy. Also, it tested with popular databases such as STARE, DRIVE, kaggle fundus images datasets are available publicly.

Deep Learning and Its Application in Diabetic Retinopathy Screening

“…The network starts with convolution blocks with activation, followed by batch normalization after each convolution layer, and all max-pooling are performed with kernel size of 3 × 3 and 2 × 2 strides. T. Chandrakumar et al [42] used DCNN to perform DR classification. With dropout introduced, additional accuracy of classification was gained.…”

Recent Innovations in Automated Detection and Classification of Diabetic Retinopathy

“…The basic steps involved in the image pre-processing is re-sizing the images. Before input the images into the model for classification, the images should convert into gray-scale [17]. Li et al, [37] proposed in their research work three steps in preprocessing.…”

“…The final step is eliminating the margin by cutting 10% from the border of the fundus images. Chandrakumar et al [17]…”

“…This study showed that using entropy images achieve 86.10%of accuracy, 73.24% of sensitivity, and 93.81% of specificity, whereas achieve 81.80% of accuracy, 68.36%of sensitivity, and 89.87%of specificity while using original fundus images. They established that transformed fundus photographs to entropy imaging can enhance the performance of deep learning and properly detect the referable DR. Chandrakumar et al [17] proposed Deep Convolutional Neural Network (DCNN) solution that find better way to classifying the retinal images with little preprocessing techniques and might give high accuracy in DR classification. their proposed approach used available public databases such as Kaggle, DRIVE, STARE and achieved around 94% of accuracy for classifying DR stages.…”

Zoom-in-Net: Deep Mining Lesions for Diabetic Retinopathy Detection

“…Early approaches [6,2] use hand-crafted features to represent the images, of which the main bottlenecks are the limited expressive power of the features. Recently, CNN based methods [4,3,9] have dramatically improved the performance of DR detection. Most of them treat CNN as a black box, which lacks intuitive explanation.…”