numpy 将数组宽度调整为零

Question

I'm a typical user of R, but in python I'm stuck.

I have a lot of images saved as NumPy array I need to resize the pad of array/images to 4k resolution from different widths which oscillated between 1620 to 2800 , the height is constant: 2160 .

I need to resize the pad of each array/image to 3840*2160 , ie. add a black border on right and left side, so that the array/image itself remains unchanged.

For resizing I try this, but the code adds black edges to all sides.

arr = np.array([[1,1,1],[1,1,1],[1,1,1],[1,1,1]])
FinalWidth = 20 

def pad_with(vector, pad_width, iaxis, kwargs):
    pad_value = kwargs.get('padder', 0)
    vector[:pad_width[0]] = pad_value

arr2 = np.pad(arr,FinalWidth/2,pad_with)

Answer 1

I think you just need hstack , assuming you want half the width to go on either side:

def pad_with(vector, pad_width):
    temp = np.hstack((np.zeros((vector.shape[0], pad_width//2)), vector))
    return np.hstack((temp, np.zeros((vector.shape[0], pad_width//2))))
arr2 = pad_with(arr,FinalWidth)
arr2 
>>> array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 1., 1., 0., 0., 0.,
        0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 1., 1., 0., 0., 0.,
        0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 1., 1., 0., 0., 0.,
        0., 0., 0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 1., 1., 0., 0., 0.,
        0., 0., 0., 0., 0., 0., 0.]])
arr2.shape
>>> (4, 23)