AttributeError: shape | When using skimage resize with pyscreenshot.grab() as input?

Question

I'm attempting to resize an image captured using pyscreenshot.grab() to 28x28 pixels

import pyscreenshot
from skimage.transform import resize

def captureAndSubsample():
    userImage = pyscreenshot.grab(bbox=(785, 335, 1125, 675))
    userImageResized = resize(userImage, (28, 28))

This code returns:

>
File "C:\Users\brad\Desktop\Development\Code\main.py", line 178, in captureAndSubsample
    userImageResized = resize(userImage, (100, 100), 3)
  File "C:\Users\brad\Desktop\Development\Code\venv\lib\site-packages\skimage\transform\_warps.py", line 144, in resize
    image, output_shape = _preprocess_resize_output_shape(image, output_shape)
  File "C:\Users\brad\Desktop\Development\Code\venv\lib\site-packages\skimage\transform\_warps.py", line 56, in _preprocess_resize_output_shape
    input_shape = image.shape
  File "C:\Users\brad\Desktop\Development\Code\venv\lib\site-packages\PIL\Image.py", line 519, in     __getattr__
    raise AttributeError(name)
AttributeError: shape
<

I've tried changing the input image to a shape:

def captureAndSubsample():
    userImage = pyscreenshot.grab(bbox=(785, 335, 1125, 675))
    userImage = shape(userImage)
    userImageResized = resize(userImage, (100, 100)).shape(340, 340)

But this returns the same as it just returns a tuple of the dimensions, any and all help is appreciated.

Answer 1

Converting userImage to a numpy array before passing to the scikit-image function does the trick.

import pyscreenshot
import numpy as np
from skimage.transform import resize

userImage = pyscreenshot.grab(bbox=(785, 335, 1125, 675))
userImage = np.array(userImage)
userImageResized = resize(userImage, (28, 28))