Retinoblastoma is a very aggressive eye cancer found almost exclusively in young children. Leukocoria (white pupil) and misaligned eyes (strabismus) are the most common symptoms. An early diagnosis is critically important for successful treatment. Retinoblastoma is almost always fatal when left untreated. Our goal through this application is to empower every person with the ability to detect and mitigate this deadliest affliction at an early stage by the use of deep learning while relying on minimum possible input from the user( just an image of the eye) and save as many lives as possible across the world. Because every single life is precious

What it does

It uses a simple four step general process to perform the diagnosis which includes:

  1. Image Input
  2. Image Processing
  3. Convolution Neural Network Analysis
  4. Image Classification ( Diagnosed Yes or No )
  5. List of hospitals nearby treating/diagnosing the disease

How we built it

Front End is a mobile app developed using React Native. Back End is developed using Springboot, MongoDB , Python and Docker. We use Convulution Neural Network for the classification of images. Datasets were complied and to the network for training.

Challenges we ran into

  • It was very difficult to find the classified images of the disease so we created our own dataset.
  • Integrating the Neural Network Model with Real Time Application.

Accomplishments that we're proud of

Its works and we are very proud of it

What's next for Keep An Eye On

As it uses Neural Networks, it is easy to implement more features and to increase the model accuracy by adding more classified images. It can also be extended to diagnose other types of eye disease such as Diabetic eyes which has visual symptoms too.

Share this project: