-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest.py
197 lines (156 loc) · 6.24 KB
/
test.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
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
from tools.mots_tools.io import *
import datetime
import math
import os
import random
import re
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import torch.utils.data
import cv2
import matplotlib.pyplot as plt
import torchvision
from roi_align import RoIAlign
from roi_align import CropAndResize
def format_box(bbox):
return torch.Tensor([[bbox[0], bbox[1], bbox[0]+ bbox[2], bbox[1] + bbox[3]]])
class SamePad2d(nn.Module):
"""Mimics tensorflow's 'SAME' padding.
"""
def __init__(self, kernel_size, stride):
super(SamePad2d, self).__init__()
self.kernel_size = torch.nn.modules.utils._pair(kernel_size)
self.stride = torch.nn.modules.utils._pair(stride)
def forward(self, input):
in_width = input.size()[2]
in_height = input.size()[3]
out_width = math.ceil(float(in_width) / float(self.stride[0]))
out_height = math.ceil(float(in_height) / float(self.stride[1]))
pad_along_width = ((out_width - 1) * self.stride[0] +
self.kernel_size[0] - in_width)
pad_along_height = ((out_height - 1) * self.stride[1] +
self.kernel_size[1] - in_height)
pad_left = math.floor(pad_along_width / 2)
pad_top = math.floor(pad_along_height / 2)
pad_right = pad_along_width - pad_left
pad_bottom = pad_along_height - pad_top
return F.pad(input, (pad_left, pad_right, pad_top, pad_bottom), 'constant', 0)
def __repr__(self):
return self.__class__.__name__
class Bottleneck(nn.Module):
expansion = 4
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(Bottleneck, self).__init__()
self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, stride=stride)
self.bn1 = nn.BatchNorm2d(planes, eps=0.001, momentum=0.01)
self.padding2 = SamePad2d(kernel_size=3, stride=1)
self.conv2 = nn.Conv2d(planes, planes, kernel_size=3)
self.bn2 = nn.BatchNorm2d(planes, eps=0.001, momentum=0.01)
self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1)
self.bn3 = nn.BatchNorm2d(planes * 4, eps=0.001, momentum=0.01)
self.relu = nn.ReLU(inplace=True)
self.downsample = downsample
self.stride = stride
def forward(self, x):
residual = x
out = self.conv1(x)
out = self.bn1(out)
out = self.relu(out)
out = self.padding2(out)
out = self.conv2(out)
out = self.bn2(out)
out = self.relu(out)
out = self.conv3(out)
out = self.bn3(out)
if self.downsample is not None:
residual = self.downsample(x)
out += residual
out = self.relu(out)
return out
class ResNet(nn.Module):
def __init__(self, architecture, stage5=False):
super(ResNet, self).__init__()
assert architecture in ["resnet50", "resnet101"]
self.inplanes = 64
self.layers = [3, 4, {"resnet50": 6, "resnet101": 23}[architecture], 3]
self.block = Bottleneck
self.stage5 = stage5
self.C1 = nn.Sequential(
nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3),
nn.BatchNorm2d(64, eps=0.001, momentum=0.01),
nn.ReLU(inplace=True),
SamePad2d(kernel_size=3, stride=2),
nn.MaxPool2d(kernel_size=3, stride=2),
)
self.C2 = self.make_layer(self.block, 64, self.layers[0])
self.C3 = self.make_layer(self.block, 128, self.layers[1], stride=2)
self.C4 = self.make_layer(self.block, 256, self.layers[2], stride=2)
if self.stage5:
self.C5 = self.make_layer(self.block, 512, self.layers[3], stride=2)
else:
self.C5 = None
def forward(self, x):
x = self.C1(x)
x = self.C2(x)
x = self.C3(x)
x = self.C4(x)
x = self.C5(x)
return x
def stages(self):
return [self.C1, self.C2, self.C3, self.C4, self.C5]
def make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
nn.Conv2d(self.inplanes, planes * block.expansion,
kernel_size=1, stride=stride),
nn.BatchNorm2d(planes * block.expansion, eps=0.001, momentum=0.01),
)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
if __name__=='__main__':
device = torch.device("cuda:"+str(0) if torch.cuda.is_available() else "cpu")
resnet = ResNet("resnet101", stage5=True)
C1, C2, C3, C4, C5 = resnet.stages()
filePath = r"E:\Challenge\MOTSChallenge\train\instances_txt"
filename = os.path.join(filePath, "{:04}.txt".format(2))
txt = load_txt(filename)
for obj in txt[1]:
mask = rletools.decode(obj.mask)
mask=torch.from_numpy(mask)
mask=mask.float()
mask=mask[None]
mask = mask[None]
mask = mask.contiguous()
boxes = format_box(rletools.toBbox(obj.mask))
# 做好坐标比例变化
box_index = torch.tensor([0], dtype=torch.int) # index of bbox in batch
# RoIAlign layer with crop sizes:
crop_height = 196
crop_width = 84
roi_align = RoIAlign(crop_height, crop_width, 0.25)
print(mask.shape)
# make crops:
crops = roi_align(mask, boxes, box_index) # 输入必须是tensor，不能是numpy
plt.imshow(crops[0][0])
plt.show()
print(crops.shape)
# RoIAlign layer with crop sizes:
boxes = torch.Tensor([[0,0,84,196]])
print(rletools.toBbox(obj.mask))
crop_height = int(rletools.toBbox(obj.mask)[3])
crop_width = int(rletools.toBbox(obj.mask)[2])
roi_align = RoIAlign(crop_height, crop_width)
crops = roi_align(crops.clone(), boxes, box_index) # 输入必须是tensor，不能是numpy
# plt.imshow(img[0][0])
# plt.show()
plt.imshow(crops[0][0])
plt.show()
break