pytorch学习笔记-常见的transform

PIL	Image.open()
tensor	ToTensor()
narrays	CV.imread()

ToTensor

输入：

from PIL import Image
# noinspection PyUnresolvedReferences
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms
writer=SummaryWriter('logs')
img=Image.open('D:\\a\\dataset\\000.png')
print(img)
# ToTensor的使用
trans_tensor=transforms.ToTensor()
img_tensor=trans_tensor(img)
writer.add_image('ToTensor',img_tensor)

输出：

 

 归一化

注：输入的图片必须是tensor类型的

公式：output[channel] = (input[channel] - mean[channel]) / std[channel]

（input-0.5）/0.5=2*input-1 ,若input[0,1]，则result[-1,1]

输入：

print(img_tensor[0][0][0])
transnorm=transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
img_norm=transnorm(img_tensor)
print(img_norm[0][0][0])
writer.add_image('Normalize',img_norm)
writer.close()

输出：

tensor(0.)
tensor(-1.)

 resize（原尺寸512*512）

输入：

print(img.size)
trans_resize=transforms.Resize((256,256))
# img pil-->resize-->img_resize pil
img_resize=trans_resize(img)
#img_resize PIL-->totensor-->img_resize tensor
img_resize=trans_tensor(img_resize)
writer.add_image("resize",img_resize,0)
print(img_resize)

输出：