The problem SHADES solves

As we all know that the world is permanently moving in online mode by all possible ways and after the end of this pandemic people will prefer 'Go online' in every walk of life but this new era comes with many problems which need to be solved.

In the upcoming scenario of 'the new normal', everything is about to get digitalized. We would like you to propose a solution for the medical industry to promote touchless diagnosis.

Our goal was to build a framework to analyze the skin pigmentation of a patient and generate a detailed medical report for the same. We have used machine learning to study the skin color of the patient.

Challenges I ran into

We broke up the problem statement into task and tried to focus on one at a time namely:

Getting relevant data for the ML model
Training and Testing the model

Our first challenge was to collect relevant data to train the model on!

We looked up research papers, did searches across the Internet until we stumbled across this one great dataset HAM10000 ("Human Against Machine with 10000 training images").

About the dataset: It contains collected dermatoscopic images from different populations, acquired and stored by different modalities. The final dataset consists of 10015 dermatoscopic images which can serve as a training set for academic machine learning purposes.

Our next big challenge: How to get it to our Users?

Not everyone can run an ML model duh!

So, to get it to our users we needed to tap into the increased mobile penetration by smartphones.

We decided to go with a mobile app because it helps us access features like the camera on devices, is handy for diagnosis and with Flutter we can target both Android, iOS, and even Web with the same codebase if we need to scale up.

The migration of TenorFlow model to first TensorFlow Lite and then the mobile device was a challenge in itself.