Chest-X-Ray-Classification.
Pytorch implementation of Chest-X-ray-Classification
Dataset
Covid-19 Chest X-Ray
Chest X-Ray images
Clasess:
- Normal
- Bacteria
- Covid-19
Set | #Normal | #Bacteria | #Covid-19 |
---|---|---|---|
#images | 1349 | 2538 | 156 |
Installations
pip install torch, torchvision, tensorflow, sklearn
Imbalanced Data (Under sampling, Over sampling, Class weights):
Data Augmentation ('noise', 'RandomRotation', 'RandomHorizontalFlip', 'brightness', 'contrast', 'RandomResizedCrop'):
'Noise'
transforms.Compose([transforms.ToTensor(), AddGaussianNoise(0.1, 0.08)])
'RandomRotation', 5%
transforms.Compose([transforms.ToTensor(), transforms.RandomRotation(degrees=5)])
'RandomHorizontalFlip', 90% probability
transforms.Compose([transforms.ToTensor(), transforms.RandomHorizontalFlip(p=0.9)])
'brightness', 50%
transforms.Compose([transforms.ToTensor(), transforms.ColorJitter(brightness=0.5)])
'contrast', 50%
transforms.Compose([transforms.ToTensor(), transforms.ColorJitter(contrast=0.5)])
'RandomResizedCrop', original size=28X28 pixels
transforms.Compose([transforms.ToTensor(), transforms.RandomResizedCrop(size=(28, 28))])