预处理之随机裁剪
预处理之随机裁剪
预处理之随机裁剪
补充:transform.invert 预处理逆操作
from PIL import Image
from torchvision import transforms
import torch
import numpy as np
def transform_invert(img_, transform_train):
"""
将data 进行反transfrom操作
:param img_: tensor
:param transform_train: torchvision.transforms
:return: PIL image
"""
if 'Normalize' in str(transform_train):
norm_transform = list(filter(lambda x: isinstance(x, transforms.Normalize), transform_train.transforms))
mean = torch.tensor(norm_transform[0].mean, dtype=img_.dtype, device=img_.device)
std = torch.tensor(norm_transform[0].std, dtype=img_.dtype, device=img_.device)
img_.mul_(std[:, None, None]).add_(mean[:, None, None])
img_ = img_.transpose(0, 2).transpose(0, 1)
if 'ToTensor' in str(transform_train):
img_ = np.array(img_) * 255
if img_.shape[2] == 3:
img_ = Image.fromarray(img_.astype('uint8')).convert('RGB')
elif img_.shape[2] == 1:
img_ = Image.fromarray(img_.astype('uint8').squeeze())
else:
raise Exception("Invalid img shape, expected 1 or 3 in axis 2, but got {}!".format(img_.shape[2]))
return img_
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.ToTensor()])
img_tensor = img_transform(img)
print(img_tensor)
print(img_tensor.shape)
img = transform_invert(img_tensor, img_transform)
img.show()
一 中心裁剪:transforms.CenterCrop
功能:从图像中心裁剪图片
主要参数说明:
- size :所需裁剪图片尺寸
原图(resize 300*300
大小)
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.ToTensor()
])
img_tensor = img_transform(img)
img = transform_invert(img_tensor, img_transform)
img.show()
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.CenterCrop(130),
transforms.ToTensor()
])
img_tensor = img_transform(img)
img = transform_invert(img_tensor, img_transform)
img.show()
中心裁剪后的效果图
二 随机裁剪:transforms. RandomCrop
功能:从图片中随机裁剪出尺寸为size的图片,如果随机超出图像范围,则会进行填充至size大小
主要参数说明:
- size :所需裁剪图片尺寸
- padding :设置填充大小
当为a时,上下左右均填充a个像素
当为(a, b)时,上下填充b个像素,左右填充a个像素
当为(a, b, c, d)时,左,上,右,下分别填充a, b, c, d - pad_if_need若图像小于设定size,则填充
- padding_ mode :填充模式,有4种模式
1、constant :像素值由fill设定
2、edge:像素值由图像边缘像素决定
3、reflect :镜像填充,最后一个像素不镜像,eg : [1,2,3,4] > [3,2,1.2,3,4,3,2]
4、symmetric :镜像填充,最后一个像素镜像,eg : [1.2,3,4] > [2,1.1,2,3,4,4.3] - fill : padding_ mode 是constant时, 设置填充的像素值
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.RandomCrop(100,100),
transforms.ToTensor()
])
img_tensor = img_transform(img)
img = transform_invert(img_tensor, img_transform)
img.show()
三 随机长宽比裁剪: Random ResizedCrop
功能:随机大小、长宽比裁剪图片
主要参数说明:
- size:所需裁剪图片尺寸
- scale:随机裁剪面积比例,默认(0.08, 1)
- ratio :机长宽比,默认(3/4, 4/3)
- interpolation:插值方法
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.RandomResizedCrop(size=200,scale=(0.5,1),ratio=(1/18,2/18)),
transforms.ToTensor()
])
img_tensor = img_transform(img)
img = transform_invert(img_tensor, img_transform)
img.show()
FiveCrop
功能:在图像的上下左右以及中心裁剪出尺寸为size的5张图片
主要参数说明:
- size :所需裁剪图片尺寸
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.FiveCrop(size=50),
])
img_tensor = img_transform(img)
for img_t in img_tensor:
img_t.show()
TenCrop
功能:在图像的上下左右以及中心裁剪出尺寸为size的5张图片,TenCrop对这5张图片进行水平或者垂直镜像获得10张图片
- size :所需裁剪图片尺寸
- vertical_ flip :是否垂直翻转
from PIL import Image
from torchvision import transforms
from utils import transform_invert
if __name__ == '__main__':
img = Image.open(r"./test.jpg").convert('RGB')
img_transform = transforms.Compose([transforms.Resize((300,300)),
transforms.TenCrop(size=50),
])
img_tensor = img_transform(img)
for img_t in img_tensor:
img_t.show()
transforms.Lambda(lambda crops: torch.stack([(transforms.ToTensor()(crop)) for crop in crops]))