Python matrix transpose Transpose

When you read a picture, you use a code like this:

Image_vector_len = Np.prod (image_size) #总元素大小, 3*55*47

img = image.open (path) arr_img = Np.asarray (img, dtype= ' float64 ') arr_img = Arr_img.transpose (2,0,1). Resha PE ((Image_vector_len,)) # 47 rows, 55 columns, each point has 3 element RGB. And put these elements in a row.

What is the consciousness of transpose? See the following example:

arr1 = Array ([[[[0, 1, 2, 3],

[4, 5, 6, 7]],

[[8, 9, 10, 11],
[12, 13, 14, 15]])
This is the original matrix. If you transpose it, execute arr2 = Arr1.transpose ((1,0,2))
Get: Array ([[[[[0], 1, 2, 3],
[8, 9, 10, 11]],

[[4, 5, 6, 7],
[12, 13, 14, 15]])

What is the process?

Arr1.shape should be (2, 2, 4) meaning 2-dimensional, 2*4 matrix

Arr1.transpose (*args) inside the parameters, so to understand, he is the order of exchange Arr1.shape, we come to the arr1.shape labeled the Horn Kazakhstan, (2[0], 2[1], 4[2]) [] is the index of shape, right,
Transpose ((1, 0, 2)) means to re-set shape in this order (2[1], 2[0], 4[2])

Although it seems that the shape before and after the transformation is 2,2,4, the problem is that transpose is the transpose
The shape is reset in the order (1,0,2), and all elements in the array are re-formed with the new matrix.

For example, 8 in the ARR1 index is (1, 0, 0) then according to the change rule, that is (0, 1, 0) look at the same position as your result arr2, and so on.

Python matrix transpose Transpose