-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtransform_gc.py
108 lines (87 loc) · 4.5 KB
/
transform_gc.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
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
import torchvision
import random
from PIL import Image, ImageOps
import numpy as np
import numbers
import math
import torch
class GroupScale(object):
"""
图片缩放器
resieze,双线性插值"""
def __init__(self, size, interpolation = Image.BILINEAR):
self.worker = torchvision.transforms.Scale(size, interpolation)
def __call__(self, img_group):
return (self.worker(img) for img in img_group)
class GroupOverSample(object):
"""进行重复采样的crop操作(随机采样和水平翻转)，最终一张图像得到10张crop的结果
对输入的n张图像都做torchvision.transforms.Scale操作，也就是resize到指定尺寸
Rescales the input PIL.Image to the given 'size'.
'size' will be the size of the smaller edge.
For example, if height > width, then image will be
rescaled to (size * height / width, size)
size: size of the smaller edge
interpolation: Default: PIL.Image.BILINEAR
"""
def __init__(self, crop_size, scale_size = None):
self.crop_size = crop_size if not isinstance(crop_size, int) else (crop_size, crop_size)
if scale_size is not None:
self.scale_worker = GroupScale(scale_size)
else:
self.scale_worker = None
def __call__(self, img_group):
if self.scale_worker is not None:
img_group = self.scale_worker(img_group)
image_w, image_h = img_group[0].size
crop_w, crop_h = self.crop_size
# 返回的offsets是一个长度为5的列表，每个值都是一个tuple，其中前4个是四个点坐标，最后一个是中心点坐标，目的是以这5个点为左上角坐标时可以在原图的四个角和中心部分crop出指定尺寸的图
# fill_fix_offset是类的静态方法
offsets = GroupMultiScaleCrop.fill_fix_offset(False, image_w, image_h, crop_w, crop_h)
oversample_group = list()
for o_w, o_h in offsets:
normal_group = list()
flip_group = list()
for i, img in enumerate(img_group):
# 按照crop_w*crop_h的大小去crop原图像，这里采用的是224*224
crop = img.crop((ow, o_h, ow + crop_w, o_h + crop_h))
normal_group.append(crop)
# 对crop得到的图像做左右翻转。最后把未翻转的和翻转后的列表合并，这样一张输入图像就可以得到10张输出了（5张crop，5张crop加翻转）
# 采用角裁剪（corner cropping）
# 这就是论文中说的corner crop，而且是4个corner和1个center。
flip_crop = crop.copy().transpose(Image.FLIP_LEFT_RIGHT)
if img.mode == 'L' and i % 2 == 0:
flip_group.append(ImageOps.invert(flip_crop))
else:
flip_group.append(flip_crop)
oversample_group.extend(normal_group)
oversample_group.extend(flip_group)
return oversample_group
class GroupMultiScanleCrop(object):
"""docstring for GroupMultiScanleCrop"""
def __init__(self, input_size, scales = None, max_distort = 1, fix_crop = True, more_fix_crop = True):
self.scales = scales if scales is not None else [1, 875, .75, .66]
self.max_distort = max_distort
self.fix_crop = fix_crop
self.more_fix_crop = more_fix_crop
self.input_size = input_size if not isinstance(input_size, int) else [input_size, input_size]
self.interpolation = Image.BILINEAR
@staticmethod
def fill_fix_offset(more_fix_crop, image_w, image_h, crop_w, crop_h):
w_step = (image_w - crop_w) // 4
h_step = (image_h - crop_h) // 4
ret = list()
ret.append((0, 0)) # upper left
ret.append((4 * w_step, 0)) # upper right
ret.append((0, 4 * h_step)) # lower left
ret.append((4 * w_step, 4 * h_step)) # lower right
ret.append((2 * w_step, 2 * h_step)) # center
if more_fix_crop:
ret.append((0, 2 * h_step)) # center left
ret.append((4 * w_step, 2 * h_step)) # center right
ret.append((2 * w_step, 4 * h_step)) # lower center
ret.append((2 * w_step, 0 * h_step)) # upper center
ret.append((1 * w_step, 1 * h_step)) # upper left quarter
ret.append((3 * w_step, 1 * h_step)) # upper right quarter
ret.append((1 * w_step, 3 * h_step)) # lower left quarter
ret.append((3 * w_step, 3 * h_step)) # lower righ quarter
return ret