How to run experiment:
- Go to experiments folder
- 3 files for segmentation (flood_segmentation.py, unet_segmentation.py, kmeans_segmentation.py )
- Choose the file you need for segmentation type to obtain segmented images
- Segmented images are now ready to be classified
- Test several DL and ML models to classify images and obtain results (classification.py)
Run WebApp (Using flask) to run experiment on single image using best model (classifies images using CNN):
- On spyder, run app.py
- Go to 127.0.0.1:5000/
- Choose an x-ray image
- To check the behaviour of model against filters, choose a certain filter
- Submit x-ray and filter to display result
For training U-net segmentation model on masks:
- Download the CXR_png, masks and test folders from Lung segmentation from Chest X-Ray dataset
- Copy the folders (CXR_png, masks, test) under \Experiments\data\unsegmented + unet masks\
For Data analysis and evaluation:
- Download the Normal and Covid data without or without masks from COVID-19 Radiography Database
- Copy the folders (COVID, Normal) under \Experiments\data\unsegmented\