Classification and detection of diabetic retinopathy using Machine learning.

Project Title

Classification and detection of diabetic retinopathy using Machine learning.

Project Area of Specialization Artificial IntelligenceProject Summary

Many patients suffer from a disease known as diabetic retinopathy which is mostly due to high blood sugar. The primary goal of the project is to classify patients having diabetic retinopathy or not, given any High-Resolution Fundus Image of the Retina. For this purpose we first do the initial image conversion which includes conversion of colored (RGB) images into perfect grey scale and then resizing it. Then we apply deep leaning, in which we fed images into convolutional neural network and then will predict whether the patient is diabetic or not. This resolution is applied on multiple images of retina. The results, so obtained are a 100 % predictive accuracy and a Sensitivity of 100 % also. Such an Automated System can easily classify images of the retina among Diabetic and Healthy patients, reducing the number of reviews of doctors.

Project Objectives

As there is a lot of complication attached with diabetic retinopathy, so there is need to develop a proper method for its detection. Although, we already have methods for the treatment of diabetic retinopathy like Focal laser treatment, Scatter laser treatment and Vitrectomy. Surgery often degrades the development of diabetic retinopathy, but is not considered as a complete cure. As it is considered as lifelong com plication, there is need to introduce a method for its diagnosis. We have diagnosis methods as well like Fluorescein angiography and Optical coherence tomography but they involve an external fluid or dye which must be applied to patient eye after retinal image is taken. But automated system which can directly whether the patient is diabetic or not is more comfortable to doctors and patients. it is time efficient as well.

Project Implementation Method

First of all we will insert all the related images. Then the images will be procssed by the image processing process. Then we will use feature extraction. Then we will classify whether the patient is having diabetic retinopathy or not. If he is having diabetic retionopathy  then we will find whther it is mild or severe.

So the methodology is following

1) Data Set

2) Data Pre-Processing

3) CNN Architecture:

in CNN We will use the following

a) Convolutional Layer

b) RELU layer

c) Max Pooling Layer

d) Flattening Layer

e) Dense Layer or Artificial Neural Network Layer

iIn this way will train the model and our project and thatt will be our methodology.

Benefits of the Project

Following are the benefits of our project.

1) It will be efficient.

2) it will be time time saving

3) It will ease the work.

4) As computer will be involved, there will be less chances of mistake.

5) It will be beneficial both for patient and Docter.

6) The dtection will be fast and efficient.

7) The degree of human Error will be less

8) It will be cost saving as well

Technical Details of Final Deliverable

Our project is based on machine learning. In machine learning , we are interested in deep learning. In deep learning we are working on CNN.

By CNN we will classify whether the Patient has Diabetic retinopathy or not. If he has then it is mild or severe. So by finding the severity of the case , we can proceed along the best available treatment.

Final Deliverable of the Project Software SystemCore Industry MedicalOther Industries Health Core Technology Artificial Intelligence(AI)Other Technologies NeuroTechSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	0
Smart Devices and Screens	Equipment	1	0	0
Devices and Screens	Equipment	1	0	0
Stationary	Equipment	2	0	0