Fixels is an application that aims to give quick diabetic retinopathy prediction based on retinal images. We trained a convolutional neural network on top of kaggle public datasets and messidor private datasets. The trained model then became nothing but the heart of our application.
In fixels, we use a python server with all the machine learning related modules including keras and tensorflow to train our model and generate prediction based on that model. For the client-side, it's a simple web application built using react and typescript for now. We are planning to develop the mobile version in the future.