X-ray Diagnosis for Covid-19

Table of Content

• Demo
• Overview
• Framework Used
• Data Collection
• Deep Learning Model
• Result

Demo

Overview

• This is a university group project about Artificial Intelligence. The CNN model takes an X-ray image and predicts if the image belongs to 1 of 3 classes with a probability: Normal, Covid, or Viral pneumonia.

Framework Used

• Keras
• Flask

Data Collection

• The dataset used takes 3000 images of the 3 classes: normal, COVID, and viral penumonia, respectively.
  • Train set contains 64 %
  • Validation set contains 16%
  • Test scontains 20%
• Source: https://www.kaggle.com/datasets/tawsifurrahman/covid19-radiography-database

Deep Learning Model

For this problem, a convolutional neural network was used to train the model. The model receives a raw image, then proceeds to clean it. After that, the cleaned image is fed into the CNN model, which proceeds to train and, in the end, classify the image with a probability for each class.

Result

The result was on the test set with 94.5% accuracy.