-
Notifications
You must be signed in to change notification settings - Fork 57
/
image_to_patches.py
73 lines (61 loc) · 2.12 KB
/
image_to_patches.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
import numpy as np
def image2patches_naive(im, patch_h, patch_w):
"""
:cut the image into patches
:param im:
:param patch_h:
:param patch_w:
:return:
"""
im_h, im_w = im.shape[0], im.shape[1]
patch_table = np.zeros((im_h - patch_h + 1, im_w - patch_w + 1, patch_h, patch_w), dtype=np.float64)
for i in range(im_h - patch_h + 1):
for j in range(im_w - patch_w + 1):
patch_table[i][j] = im[i:i + patch_h, j:j + patch_w]
return patch_table
def image2patches(im, patch_h, patch_w):
"""
:cut the image into patches
:param im:
:param patch_h:
:param patch_w:
:return:
"""
im_h, im_w = im.shape[0], im.shape[1]
im_h_idx = np.arange(im_h - patch_h + 1)
im_h_idx = im_h_idx[:, np.newaxis, np.newaxis, np.newaxis]
im_w_idx = np.arange(im_w - patch_w + 1)
im_w_idx = im_w_idx[np.newaxis, :, np.newaxis, np.newaxis]
patch_h_idx = np.arange(patch_h)
patch_h_idx = patch_h_idx[np.newaxis, np.newaxis, :, np.newaxis]
patch_w_idx = np.arange(patch_w)
patch_w_idx = patch_w_idx[np.newaxis, np.newaxis, np.newaxis, :]
h_idx = im_h_idx + patch_h_idx
w_idx = im_w_idx + patch_w_idx
return im[h_idx, w_idx].astype(np.float64)
if __name__ == '__main__':
import cv2
from utils import ind_initialize
im = cv2.imread('test_data/image/Cameraman.png', cv2.IMREAD_GRAYSCALE)
height, width = im.shape[0], im.shape[1]
k = 8
n = 16
p = 3
# row_ind = ind_initialize(height - k + 1, n, p)
# column_ind = ind_initialize(width - k + 1, n, p)
# im = cv2.resize(im, (512, 512))
# for i in range(100):
# im[i, i] = 0
# im[i, i + 50] = 0
# im[i, i + 100] = 0
# im[i, i + 150] = 0
# cv2.imshow('im', im)
patch_size = 3
res = image2patches_naive(im, patch_size, patch_size)
res1 = image2patches(im, patch_size, patch_size)
print(np.max(res - res1))
# res = image2patches_v1(im, k, 0)
# for i in range(res.shape[0]):
# for j in range(res.shape[1]):
# cv2.imshow('patches', res[i, j].astype(np.uint8))
# cv2.waitKey(100)