EYECURE

Project Title

EYECURE

Project Area of Specialization Artificial IntelligenceProject Summary

Summary:

Eyecure is AI based system trained to detect diabetic retinopathy ,hypertensive retinopathy and macular degeneration. These diseases affects blood vessels in retina that lines the back of eye. These diseases are common cause of vision loss among people with diabetes and leading cause of vision impairment and blindness among working age adults.

The proposed system is an integrated software hardware solution where first a machine learning model is trained by providing large number of disease pattern images.

The system will take retinal fundus image of the patient as an input, process it using filtering methods to remove noise and environmental interference from image and the system will then classify the input image and predict the disease by using existing machine learned model.

Project Objectives

Objectives:

To predict the possibilities of retinal diseases a patient can have using blood vessel extraction.

Project Implementation Method

Phase 1

Analysis/Surveying

Phase 2

image processing component

Phase 3

machine learning

Phase 4 

Implementation Of Tensor Flow & OpenCV

Phase 5

Developing Client Server App

Phase 6 

Testing

Benefits of the Project

Benefits:

• Provide early diagnoses of diseases.
• Prevent misdiagnoses.
• Cost-Effective.

Technical Details of Final Deliverable

Technical implementation:

Implementation is divided into following components:

Input:

Retinal image of the patient.

Image processing:

• Resize image for easier computation
• Convert RGB to GrayScale
• Contrast Enhancement of gray image using CLAHE(Contrast Limited Adaptive Histrogram Equalization)
• Background Exclusion:

It is the step in which the background variations is eliminated so that the foreground object be more easily analyzed.

Machine learning:

The machine learning will be done using Scikit-learn built in python library that has various classifications which is used for training and testing of our system and for validation of the system we used technique, called ISO data that provide an automated threshold value for binarization of the given data.

Some of the datasets which we used are

• DRIVE: That has 40 fundus images divided into 2 sets a test set and a training sets
• STARE: The dataset contain total 20 eye fundus images with a resolution of 700*605

Output:

The system will use  matplotlib

 library of python to generate graph or histogram that shows the predicted diseases.

Final Deliverable of the Project HW/SW integrated systemCore Industry MedicalOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable 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)	70000
Fundus Camera	Equipment	1	70000	70000