CoronoVirus-Image-Project

Approach

Download and merge the following data sets

1. Covid 19 X-Ray DataSet
2. Normal Chest X-Ray Images
3. Bacterial and Viral Pneumonia Chest X-Ray Images

After downloading and merging the datasets we need to load these images into Pytorch for Deep Learning.

We also need to consider what views we are considering. Since the Kaggle Data Set

To load these images into Pytorch

After loading these images lets get these images to Grayscale so that instead of 3 channels there is only one(i.e White and Black)

We also need to perform Pytorch transformations. The following are the transformations we need to perform:

• Centering - We can Center all our images using
• Resizing - Since we have images from different sources, its important to have them resized to the same amount, we will create images of 224x224

We need to first convert the raw data into Pytorch Tensors these tensors need to be split into X_train, y_train and X_test and y_test.

• X_train - contains a list of images meant for training
• y_train - contains a list of labels for train images
• X_test - contains a lisf of images meant for testing
• y_test - contains a list of labels for test images

We might also want to have some sort of validation data to try and improve our model.

• X_train - contains a list of images to perform validation
• y_train - contains a list of labels for the validation images

Reading Covid 19 Chest X-ray

Before we go throug the Covid 19 X ray data set

We need to understand the scheme documented here

Libraries

1. TorchXrayVision - an open source software to work with any chest xray datasets
2. Pandas
3. Pytorch
4. Pandas
5. OpenCV

Tools

1. Google Colab