关于python：PIL：Image.fromarray(img.astype(‘uint8′)，mode =’RGB’)返回灰度图像


PIL: Image.fromarray(img.astype('uint8'), mode='RGB') returns grayscale image

我已经将大小为torch.Size([3, 28, 28])的pytorch张量转换为大小为(28, 28, 3)的numpy数组，并且似乎没有任何问题。然后，我尝试使用img = Image.fromarray(img.astype('uint8'), mode='RGB')将此图像转换为PIL图像，但是当我期望它为(28, 28, 3)(或(3, 28, 28))时，返回的img的尺寸为(28, 28)。我不明白为什么会这样。我确保转换为uint8并使用RGB模式，就像其他海报在网上建议的那样，但是这些海报(也不使用np.ascontiguousarray)都没有帮助。

PIL版本1.1.7

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# This code implements the __getitem__ function for a child class of datasets.MNIST in pytorch
# https://pytorch.org/docs/stable/_modules/torchvision/datasets/mnist.html#MNIST

img, label = self.data[index], self.targets[index]
assert img.shape == (3, 28, 28), \\
       (f'[Before PIL] Incorrect image shape: expecting (3, 28, 28),'
        f'received {img.shape}')

print('Before reshape:', img.shape)    # torch.Size([3, 28, 28])
img = img.numpy().reshape(3, 28, 28)
img = np.stack([img[0,:,:], img[1,:,:], img[2,:,:]], axis=2)
print('After reshape:', img.shape)     # (28, 28, 3)

# doing this so that it is consistent with all other datasets
# to return a PIL Image
img = Image.fromarray(img.astype('uint8'), mode='RGB') # Returns 28 x 28 image

assert img.size == (3, 28, 28), \\
       (f'[Before Transform] Incorrect image shape: expecting (3, 28, 28), '
        f'received {img.size}')

编辑：这是一个最小的示例。如果有任何帮助，我将保留上述内容。

1
2
3
4
5
6
7
8
9
from PIL import Image
import numpy as np

img = np.random.randn(28, 28, 3)
img = Image.fromarray(img.astype('uint8'), mode='RGB') # Returns 28 x 28 image

assert img.size == (28, 28, 3), \\
       (f'[Before Transform] Incorrect image shape: expecting (3, 28, 28), '
        f'received {img.size}')

AssertionError: [Before Transform] Incorrect image shape: expecting (3, 28, 28), received (28, 28)

我想您想要这样，其中RGB值的范围是0..255：

范围内的整数

1
2
3
4
5
6
7
8
import numpy as np
from PIL import Image

# Make random 28x28 RGB image
img =np.random.randint(0,256,(28,28,3), dtype=np.uint8)

# Convert to PIL Image
pImg=Image.fromarray(img, mode='RGB')

现在检查我们拥有什么：

1
2
In [19]: pImg                                                                                      
Out[19]: <PIL.Image.Image image mode=RGB size=28x28 at 0x120CE9CF8>

并保存：

1
pImg.save('result.png')

enter image description here

相关讨论