GUI-fying to Build Deep Learning Model for Classification Problem

Build Detection and Diagnosis Tools of Covid-19 on Chest Radiographs. Everything is in GUI.

Author

Kevin Chng

Overview

This example provides the workflow how to build deep learning model using GUI applications in MATLAB. I was inspired from a MathWork's blog posted by Johanna Pingel (Link : Covid-19 Medical image Detection). It describes how to build detection and diagnosis tools of COVID-19 radiography image, it offers a valueble second opinion to the doctors and assit them in the screening process. This type of mechanism would also assist in providing results to the doctors quickly.
In my example, we are fying through the GUI to build a deep learning (DL) model to detect COVID-19 on Chest Radiography using MATLAB.

Dataset

1) COVID-19 Radiography Medical Image, I randomly extract 90 images from COVID-19 dataset curated by Dr. Joseph Cohen.
2) Normal Chest Radiography Medical Image, I reandomly extract 90 images from Kaggle dataset posted by Paul Mooney.

Highlights

1) Understand the concept of AI and Transfer Learning
2) Learn how to train Classificaion Model using Transfer Learning
3) Learn how to build deep learning model using Deep Network Designer
4) Learn how to fine tune the deep learning model using Experiment Manager

Product Focus

MATLAB (R2020a)
Deep Learning Toolbox (R2020a)
If you are using MATLAB version R2019a/R2019b, you may consider to use transfer learning app (Community App)

Workflow