python - 瞄准质心 - scipy/numpy

Question

鉴于以下情况:

from scipy.ndimage import center_of_mass
from numpy import array

A = array([
  [ 255, 255, 0, 0 ]
  [ 255, 255, 0, 0 ],
  [ 0   , 0,  0, 0 ],
  [ 0   , 0,  0, 0 ]
])

cm = center_of_mass(A)
# cm = (0.5, 0.5)

centered = ???
cmc = center_of_mass(centered)
# cmc ~= (1.5,1.5)

我们如何移动这个 ndarray/图像，使其基于其质心居中？

我们的目标结果如下:

centered = array([
  [ 0,   0,   0, 0 ],
  [ 0, 255, 255, 0 ],
  [ 0, 255, 255, 0 ],
  [ 0,   0,   0, 0 ]
])

Answer 1

很简单:

1. 求几何中心

   import numpy as np
   
   c1 = center_of_mass(np.ones_like(A))
   #or : c1 = [A.shape[0]/2.,A.shape[1]/2.]
   

2. 按差值平移矩阵

   S = np.roll(A, c1[0]-cm[0] , axis=0)
   S = np.roll(S, c1[0]-cm[0] , axis=1)
   

答案是:

    Out[18]: 
    array([[  0,   0,   0,   0],
           [  0, 255, 255,   0],
           [  0, 255, 255,   0],
           [  0,   0,   0,   0]])