Radon Health - The Web App

Welcome to my 4th Honors experience reflection! This project focuses on developing web app for a data analytics application. The project is based on the use of frontend frame work like Flask, CSS scripts and also Deep Learning models, a technical process, to make predictions on X-rays images. The goal is to give doctors and radiologists an upper hand in the clinical analysis of the medical scans for chest infections.

Following the pandemic and will to help, today, I’m here to pitch my idea for Radiology scanning through an advanced AI platform. Over 12 million Americans are misdiagnosed with 30% or more errors in complex diagnoses and over 750$ billion dollars are wasted in healthcare. Introducing Smart Chest Radon Health, an advanced Data analytics platform that will enable radiologists to make smart and accurate decisions using state of art Neural network models. This reflection is divided into 3 sections to make the explanations easier to understand.

Normal (Left 2 columns) VS Infected (Right two columns)

Building a web app involves creating a dynamic website that allows users to interact with it and perform certain actions or tasks. Here are the basic steps involved in building a web app:

1. Choose a programming language: There are several programming languages you can use to build a web app, such as JavaScript, Python, Ruby, PHP, and more. For this project, I have used Python as the main language.

2. Choose a framework: A framework is a collection of tools and libraries that makes it easier to build web apps. Usually Flask is a compatible version when it comes to frontend apps for Python.

3. Create the basic structure of your web app: This involves creating the HTML, CSS, and JavaScript files that make up your app. I used HTML to setup page in the app and CSS to make it cleaner.

Breaking Down the language:

For this part, I created a web app directly that contained all code for the processing the images and making predictions. In simple words, when the user uploads the image of an X-ray, the flask code as a delivery person, meaning it takes the image and delivers it to the backend. Once the image goes in backend, the image is processing using the Jupyter scripts which I build in my 3rd Honors experience and the model makes a prediction. Once the prediction is made, the flask again being the delivery person, takes the prediction and displays it on website where the user sees the results. Below is a part of the flask code.

We can see how the image is captured by the code and then reshaped and resized depending on the dimensions of the user inputs. We can also see the model at the top on line 14, which basically the prediction part from 3rd Honors Experience. I talked about the delivery part, and below is the code script that does that.

Once the processing is done, the app just sends the response to the frontend server hosted on a web app. For instance, in the below picture, I uploaded an example picture to check the chances of an infection. The result was close to 99% based on the trained model.

Summary

Lastly, the main learning goal for this experience was just trying to dig deeper into the world of AI for Medical Imaging but this time through the eyes of frontend developer. Throughout the project, I learned how not only an X-ray image can tell the affected parts but also give statistical details like sensitivity and specificity numbers. At the same time, it helped develop new skills in building web apps!