-
Notifications
You must be signed in to change notification settings - Fork 0
/
cocoLoader.py
164 lines (141 loc) · 5.38 KB
/
cocoLoader.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
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
from collections import defaultdict
import numpy as np
from PIL import Image
from torch.utils.data import Dataset, DataLoader
import torchvision.transforms as transforms
import os
from pycocotools.coco import COCO
import cv2
import torch
category_dict = {'person': 1,
'bicycle': 2,
'car': 3,
'motorcycle': 4,
'airplane': 5,
'bus': 6,
'train': 7,
'boat': 9,
'bird': 16,
'bottle': 44,
'cat': 17,
'chair': 62,
'cow': 21,
'dining table': 67,
'dog': 18,
'horse': 19,
'potted plant': 64,
'sheep': 20,
'couch': 63,
'tv': 72}
temp_dict = {1: 0,
2: 1,
3: 2,
4: 3,
5: 4,
6: 5,
7: 6,
9: 7,
16: 8,
44: 9,
17: 10,
62: 11,
21: 12,
67: 13,
18: 14,
19: 15,
64: 16,
20: 17,
63: 18,
72: 19}
class CocoDataset(Dataset):
def __init__(self, rootDir: str, annFile: str, transform = None) -> None:
super(CocoDataset, self).__init__()
self.coco = COCO(annFile)
self.ids = list()
for _, val in category_dict.items():
img_ids = self.coco.getImgIds(catIds=[val])
requiredID = list()
for ids in img_ids:
anns = self.coco.loadAnns(self.coco.getAnnIds(imgIds=[ids], iscrowd=None))
target_classes = defaultdict()
for i in range(len(anns)):
if(anns[i]['category_id'] in category_dict.values()):
if(anns[i]['category_id'] in target_classes.keys()):
target_classes[anns[i]['category_id']] += 1
else:
target_classes[anns[i]['category_id']] = 1
else:
target_classes = defaultdict()
break
if(len(target_classes.keys()) == 1):
requiredID.append(ids)
self.ids.extend(requiredID)
self.ids = list(set(sorted(self.ids)))
print(len(self.ids))
self.root = rootDir
self.transform = transform
self.blur_transform = transforms.Compose([transforms.ToTensor(),transforms.Resize((256, 256))])
def motion_blur_horizontal(self,img_path,kernel_size):
img = cv2.imread(img_path)
kernel_h = np.zeros((kernel_size, kernel_size))
kernel_h[int((kernel_size - 1)/2), :] = np.ones(kernel_size)
kernel_h /= kernel_size
horizontal_mb = cv2.filter2D(img, -1, kernel_h)
horizontal_mb = cv2.cvtColor(horizontal_mb, cv2.COLOR_RGB2BGR)
return horizontal_mb
def motion_blur_vertical(self,img_path,kernel_size):
img = cv2.imread(img_path)
kernel_v = np.zeros((kernel_size, kernel_size))
kernel_v[:, int((kernel_size - 1)/2)] = np.ones(kernel_size)
kernel_v /= kernel_size
vertical_mb = cv2.filter2D(img, -1, kernel_v)
vertical_mb = cv2.cvtColor(vertical_mb, cv2.COLOR_RGB2BGR)
return vertical_mb
def gaussian_blur(self,img_path,k_size):
img = cv2.imread(img_path)
img = cv2.GaussianBlur(img,(k_size,k_size),0)
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
return img
def avg_blur(self,img_path,k_size):
img = cv2.imread(img_path)
img = cv2.blur(img, (k_size,k_size))
img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
return img
def getBlurredOutput(self,img_path):
img_path = os.path.join(self.root, img_path)
p = torch.rand(1)
k_size = 15
if p<=0.25:
return self.motion_blur_vertical(img_path,k_size)
if p<=0.5:
return self.motion_blur_horizontal(img_path,k_size)
if p<=0.75:
return self.avg_blur(img_path,k_size)
else:
return self.gaussian_blur(img_path,k_size)
def __len__(self) -> int:
return len(self.ids)
def _load_image(self, ids: int) -> Image.Image:
path = self.coco.loadImgs(ids)[0]["file_name"]
return Image.open(os.path.join(self.root, path)).convert("RGB"),path
def _load_target(self, ids: int):
target = self.coco.loadAnns(self.coco.getAnnIds(ids, iscrowd=None))
return temp_dict[target[0]['category_id']]
def __getitem__(self, index: int):
ids = self.ids[index]
image,path = self._load_image(ids)
blurredImage = self.getBlurredOutput(path)
target = self._load_target(ids)
if(self.transform is not None):
image = self.transform(image)
blurredImage = self.transform(blurredImage)
return {'image': image, 'inputImg': blurredImage, 'class': target}
if(__name__ == '__main__'):
trans = transforms.Compose([
transforms.ToTensor(),
transforms.Resize((256, 256))
])
cocoData = CocoDataset('train2014', 'annotations/instances_train2014.json', transform=trans)
dl = DataLoader(cocoData, batch_size=2, shuffle=True)
for step, (data) in enumerate(dl):
a = step