Target Size

2022-09-18
Image Resize, Image Padding, Image Scanning

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
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
def scanImageWithWindowSizeAutoResize(
image,
width,
height,
return_direction=False,
threshold=0.1,  # minimum 'fresh' area left for scanning
):  # shall you use torch? no?
shape = image.shape
assert len(shape) == 3
ih, iw, channels = shape
targetWidth = max(width, math.floor(iw * height / ih))
targetHeight = max(height, math.floor(ih * width / iw))
resized = cv2.resize(
image, (targetWidth, targetHeight), interpolation=cv2.INTER_CUBIC
)
# determine scan direction here.
imageSeries = []
if targetWidth / targetHeight == width / height:
imageSeries = [resized]  # as image series.
direction = None
elif targetWidth / targetHeight < width / height:
direction = "vertical"
# the scanning is along the vertical axis, which is the height.
index = 0
while True:
start, end = height * index, height * (index + 1)
if start < targetHeight:
if end > targetHeight:
if 1 - (end - targetHeight) / targetHeight >= threshold:
end = targetHeight
start = targetHeight - height
else:
break
# other conditions, just fine
else:
break  # must exit since nothing to scan.
cropped = resized[start:end, :, :]  # height, width, channels
imageSeries.append(cropped)
index += 1
else:
direction = "horizontal"
index = 0
while True:
start, end = width * index, width * (index + 1)
if start < targetWidth:
if end > targetWidth:
if 1 - (end - targetWidth) / targetWidth >= threshold:
end = targetWidth
start = targetWidth - width
else:
break
# other conditions, just fine
else:
break  # must exit since nothing to scan.
cropped = resized[:, start:end, :]  # height, width, channels
imageSeries.append(cropped)
index += 1
if return_direction:
return imageSeries, direction
else:
return imageSeries
def resizeImageWithPadding(
image,
width,
height,
border_type: Literal["constant_black", "replicate"] = "constant_black",
):
shape = image.shape
assert len(shape) == 3
ih, iw, channels = shape
targetWidth = min(width, math.floor(iw * height / ih))
targetHeight = min(height, math.floor(ih * width / iw))
resized = cv2.resize(
image, (targetWidth, targetHeight), interpolation=cv2.INTER_CUBIC
)
BLACK = [0] * channels
top = max(0, math.floor((height - targetHeight) / 2))
bottom = max(0, height - targetHeight - top)
left = max(0, math.floor((width - targetWidth) / 2))
right = max(0, width - targetWidth - left)
if border_type == "constant_black":
padded = cv2.copyMakeBorder(
resized, top, bottom, left, right, cv2.BORDER_CONSTANT, value=BLACK
)
elif border_type == "replicate":
padded = cv2.copyMakeBorder(
resized, top, bottom, left, right, cv2.BORDER_REPLICATE, value=BLACK
)
else:
raise Exception("unknown border_type: %s" % border_type)
return padded

Read More
▲