Diagnosis of Glaucoma by Cup-to-Disc Ratio Mobile Application

Project Title

Diagnosis of Glaucoma by Cup-to-Disc Ratio Mobile Application

Project Area of Specialization Artificial IntelligenceProject Summary

Glaucoma is an eye disease which causes irreversible loss of vision and may lead to blindness. According to the World Health Organization (WHO), glaucoma is the second leading cause of blindness worldwide after cataracts. Globally, 60.5 million had glaucoma in 2010. Given the aging of the world's population, this number may increase to almost 80 million by 2020. Mostly these patients crossed 40, and more than 50% of patients are unaware of this because of the asymptomatic and slow process of Glaucoma. It affects the visual field by which the peripheral vision of the patient is being affected, and the vision is being decreased after that he/she loses its sight. 

But if it is detected and cured at the time, it could be controlled. with a regular checkup, eye drops or operation.

so we made our project (DGCDR) to early detect and diagnose the Glaucoma by measuring CDR  with real-time images. It is an android application which takes pictures and measures the ratio of the optic disc and optic cup and suspects whether the patient has Glaucoma or not.

Project Objectives

The project is a real-time android application it would use a smartphone with high resolution to diagnose the Glaucoma which is an eye disease. It uses different algorithms and techniques to classify the image and could suspect the Glaucoma by measuring CDR. It may help the related technician in any clinic as the tools and machines to detect Glaucoma are so expensive to be afforded by a single person(technician). It also could be used in rural areas where the facilities are negligible and it cost less as compare to any workshop of Glaucoma done in rural areas.

Project Implementation Method

It uses a different machine learning algorithm and techniques to classify different images of the eye. It uses Cloud to store the database which is used for training and testing the machine. Cloud may also contain the images captured by the user for diagnosis. For the classification of images, we use python and its different libraries with the Jupyter platform. Some specifics are used as:

TensorFlow Lite Model will be used for inference in this app.

ML Kit (package or mobile SDK that brings Google's machine expertise to Android and iOS apps) will use TensorFlow lite models.

ML Kit acts as an API layer to our custom model, which would make it easy to run and use.

Beside this DGCDR needs a Smartphone with a high resolution to make it work properly. And it needs a lens between the camera and eye to take the picture of the retina of the eye.

Benefits of the Project

The DGCDR is a real-time android application project which will be used in the clinic or anywhere by a related technician. With some basic information and knowledge, an individual can also use it by following all the given instruction properly. At present, the Glaucoma is diagnosed by many expensive screening tools which measure the IOP, CDR and visual field to diagnose Glaucoma.

In this method, there is a nice amount of fees and time consuming with at least 10 – 20 minute required time. Excluding waiting time in the queue. And when it's about to test it in any clinic, the tools or very expensive and hard to afford by any doctor or technician, individually. To resolve this issue an android application is developed and this application is mainly for that clinic and the individual person with skills. It is also a better way and a good solution to be used in rural areas where there is a lack of facilities.

The doctors or technician can use it in real time to capture the image of the patient’s retina and the application will measure the ratio between the optic cup and optic disc. It is very simple and user-friendly you just have to select the eye (left or right) and capture the image in very few time you can have the result. It also contains the images captured previously or you may also be the first one to capture the image and suspect glaucoma in any clinic.

Technical Details of Final Deliverable

The product must have to give some basic information about glaucoma disease as an introducing guide to the user.

The product must have to provide the guide as instruction of how the product could be used.

The product must have the ability to connect the user to the team of developers in case of any query or issue in DGCDR application.

The product should have to take the efficient and proper picture from camera or may access the gallery images of the device it is running on.

The product must preprocess image and then classify it by measuring the ratio of the optic disc and optic cup with the maximum possible level of accuracy.

The result provided by DGCDR must be easy to understand and if any shortcodes are used in result then it must be described in the system earlier or with that code.

The product must allow the user to access all the informative and related documents of the DGCDR system.

Final Deliverable of the Project Software SystemType of Industry Health Technologies Artificial Intelligence(AI)Sustainable 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)	53160
PEEK Retina	Equipment	1	53160	53160