COVID-19 X-Ray Diagnosis

This model takes a chest X-ray image as an input and diagnoses if the patient is COVID-19 positive. The dataset used is a combination of datasets obtained from http://ugm.id/MVDataset and the "COVID-19 Radiography Database" dataset. The dataset contains x-ray images of normal people and people with COVID-19. It is calculated that the dataset contains 10,290 negative images and 3,674 positive images of COVID-19.

Figure 1 Dataset Sample

The training images are resized into 100x100 images and then fed into the prediction model. The prediction algorithm used is Convolutional Neural Network with architecture as below:

Figure 2 CNN Model Architecture

Once trained, the model is then used to make predictions. Below are some examples of the prediction results.

Figure 2 CNN Model Architecture