Question

我正在使用python进行对象检测的图像处理。我需要将我的图像分成所有可能的块。例如，鉴于这个玩具形象：

x = np.arange(25)
x = x.reshape((5, 5))

[[ 0  1  2  3  4]
 [ 5  6  7  8  9]
 [10 11 12 13 14]
 [15 16 17 18 19]
 [20 21 22 23 24]]

我想检索给定大小的所有可能块，例如2x2块是：

[[0 1]
 [5 6]]
[[1 2]
 [6 7]]

..等等。我怎么能这样做？

Answer 1

scikit图片extract_patches_2d执行此操作

>>> from sklearn.feature_extraction import image
>>> one_image = np.arange(16).reshape((4, 4))
>>> one_image
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11],
       [12, 13, 14, 15]])
>>> patches = image.extract_patches_2d(one_image, (2, 2))
>>> print(patches.shape)
(9, 2, 2)
>>> patches[0]
array([[0, 1],
       [4, 5]])
>>> patches[1]
array([[1, 2],
       [5, 6]])
>>> patches[8]
array([[10, 11],
       [14, 15]])

Answer 2

您可以使用以下内容：

def rolling_window(arr, window):
    """Very basic multi dimensional rolling window. window should be the shape of
    of the desired subarrays. Window is either a scalar or a tuple of same size
    as `arr.shape`.
    """
    shape = np.array(arr.shape*2)
    strides = np.array(arr.strides*2)
    window = np.asarray(window)
    shape[arr.ndim:] = window # new dimensions size
    shape[:arr.ndim] -= window - 1
    if np.any(shape < 1):
        raise ValueError('window size is too large')
    return np.lib.stride_tricks.as_strided(arr, shape=shape, strides=strides)

# Now:
slices = rolling_window(arr, 2)
# Slices will be 4-d not 3-d as you wanted. You can reshape
# but it may need to copy (not if you have done no slicing, etc. with the array):
slices = slices.reshape(-1,slices.shape[2:])

Answer 3

>>> x = np.arange(12).reshape(3,4)
>>> print x
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
>>> r = 2
>>> for z in [x[i:i+r,j:j+r] for i in xrange(x.shape[0] - r + 1) for j in xrange(x.shape[1] - r + 1)]:
...   print z
... 
[[0 1]
 [4 5]]
[[1 2]
 [5 6]]
[[2 3]
 [6 7]]
[[4 5]
 [8 9]]
[[ 5  6]
 [ 9 10]]
[[ 6  7]
 [10 11]]

将numpy图像数组切成块

3 个答案: