CLozy

import torchvision.transforms as transforms
from torch.utils.data import Dataset, DataLoader

#data augmentation 
train_transforms = transforms.Compose([
                                       #transforms.ToPILImage(),
                                       transforms.RandomHorizontalFlip(),
                                       transforms.RandomRotation(10),
                                       transforms.RandomResizedCrop(224),
                                       transforms.ToTensor(),

CLozy

image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')

cropped = image.crop((200,200,400,400))
cropped

CLozy

image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')

#plot original image
f = pyplot.figure(figsize=(30,30))
f.add_subplot(131)
pyplot.imshow(image)

#45 degrees rotation
f.add_subplot(132)
pyplot.imshow(image.rotate(45))

CLozy

image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')

#horizontal flip
hor_flip = image.transpose(Image.FLIP_LEFT_RIGHT)

#vertical flip
ver_flip = image.transpose(Image.FLIP_TOP_BOTTOM)

#plot all three using matplotlib
f = pyplot.figure(figsize=(30,30))

CLozy

image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')
print(image.size)

#resize image and ignore original aspect ratio
img_resized = image.resize((400,400))
print(img_resized.size)

img_resized

CLozy

#create thumbnail of an image
image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')

#report size
print(image.size)

#create thumbnail and preserve aspect ratio
image.thumbnail((400,400))

#print size

CLozy

from PIL import Image
from matplotlib import pyplot


#load image
image = Image.open('/content/gdrive/MyDrive/ComputerVisionEngineer/images/deadstar.jpg')

#summarize some details about the image
print(image.format)
print(image.mode)

CLozy

#splitting the dataset.csv file into train , test splits and saving them as csv files
from sklearn.model_selection import train_test_split
dataset_df = pd.read_csv(r'/content/gdrive/My Drive/Malaria/dataset.csv')
train,test = train_test_split(dataset_df, test_size=0.1)

train.to_csv(r'gdrive/My Drive/Malaria/train.csv', index=False)
test.to_csv(r'gdrive/My Drive/Malaria/test.csv', index=False)

CLozy

for param in model.parameters():
  param.requires_grad = False

from collections import OrderedDict
classifier = nn.Sequential(OrderedDict([
                                        ('fc1', nn.Linear(25088, 4096)),
                                        ('relu1', nn.ReLU()),
                                        ('drop1', nn.Dropout(p=0.5)),
                                        ('fc2', nn.Linear(4096, 4096)),
                                        ('relu2', nn.ReLU()),

CLozy

#data augmentation 
train_transforms = transforms.Compose([
                                       #transforms.ToPILImage(),
                                       transforms.RandomHorizontalFlip(),
                                       transforms.RandomRotation(10),
                                       transforms.RandomResizedCrop(224),
                                       transforms.ToTensor(),
                                       transforms.Normalize([0.485, 0.456, 0.406],
                                                      [0.229, 0.224, 0.225])])