Eye for Diabetes

How do you take a photograph of the back of the eye?

The images are taken with a high-quality camera, in a process called digital retina photography. The digital camera is able to produce an image within a few seconds, and provides an overview of the inside of the eye and the retina. The image helps doctors to understand what is going on behind the eye lens, and helps to identify retinal diseases, such as diabetic retinopathy. Other eyes related issues can also be identified through retina images, such as glaucoma, macular degeneration, cataract, etc. The images are often used as a baseline for further examinations.

The colours in retina images vary wildly. Why is that?

The difference in colour depends on the type of digital camera of the optometrist. A different type of camera, and also different lighting and setting at the optometrist can cause a variance in more bright or dark colours.

Should I submit a photograph to be able to participate in this research project?

We use images from a public, anonymised dataset. This set contains more than enough images for us to work with. You don't have to submit your own images.

I am worried about my eye health, can I have my eyes screened?

We are not medical professionals and we can not provide any health related recommendations. We advise you to contact your doctor in case of any health concerns. If you are suffering from diabetes, it is recommended to have your eyes screened for signs of DR once per year, even if you do not experience any symptoms!

Should I create a Zooniverse account?

You don’t have to, but it is a good idea to do so. Creating a Zooniverse account is easy, for free, and has a clear number of advantages. The Zooniverse platform allows logged in users to send and receive messages. You will get the possibility to discuss images and your home page will keep track of the number of annotations you have already done. Also, with each annotation, the Zooniverse account name of the user who did the annotation is also stored. This allows to keep track of the most active users.

Will the created data set become available to other researchers?

Yes, the data set created in this project will be made available to other researchers after the data collection has been completed and all data has been processed.

Is the EyePacs dataset not already labeled by EyePacs?

No and yes. Yes: each fundus photo of the EyePacs dataset has been labeled by EyePacs with the corresponding DR class. And no: the individual lesions have not been labeled on the EyePacs images. In other words: we know at which stage diabetic retinopathy is in each image, but the images have not been annotated like we are asking you to do in this project.
The existing DR class labels are in our view not sufficient because recent research has proven DR class labels to be very inconsistent. Localizing and quantifying all relevant lesions is a difficult task, even for specialists, but is required to determine the correct DR grade. By marking all lesions we expect to obtain a more consistent DR scoring.

Who are the Eye for Diabetes participants?

We are having participants from different ages ranges all over the world (mostly from Belgium, the United Kingdom and the United States of America). Most of our participants have experience with other labelling projects on the Zooniverse platform, and are having a keen interest in science. Some of our participants are feeling personally connected towards the topic of eye health, since they or family members have diabetes. Would you like to get to know some of them? Click the following link to read testimonials of our top contributors. We want to thank you all for your help!