-
Notifications
You must be signed in to change notification settings - Fork 0
/
time_obtain_CAM_masking.py
206 lines (187 loc) · 10.4 KB
/
time_obtain_CAM_masking.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
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
import torch
from torch import multiprocessing, cuda
from torch.utils.data import DataLoader
import torch.nn.functional as F
from torch.backends import cudnn
import time
import numpy as np
import importlib
import argparse
import os
os.environ["CUDA_VISIBLE_DEVICES"] = '1'
from numpy.linalg import lstsq
from scipy.linalg import orth
import voc12.dataloader
from misc import torchutils, imutils
import cv2
from gradCAM import GradCAM
cudnn.enabled = True
parser = argparse.ArgumentParser()
# Environment
parser.add_argument("--num_workers", default=os.cpu_count()//2, type=int)
parser.add_argument("--voc12_root", default='Dataset/VOC2012_SEG_AUG/', type=str,
help="Path to VOC 2012 Devkit, must contain ./JPEGImages as subdirectory.")
# Dataset
parser.add_argument("--train_list", default="voc12/train.txt", type=str)
parser.add_argument("--val_list", default="voc12/val.txt", type=str)
parser.add_argument("--infer_list", default="voc12/train.txt", type=str,
help="voc12/train_aug.txt to train a fully supervised model, "
"voc12/train.txt or voc12/val.txt to quickly check the quality of the labels.")
parser.add_argument("--chainer_eval_set", default="train", type=str)
# Class Activation Map
parser.add_argument("--cam_network", default="net.resnet50_cam", type=str)
parser.add_argument("--cam_crop_size", default=512, type=int)
parser.add_argument("--cam_batch_size", default=2, type=int) # original: 16
parser.add_argument("--cam_num_epoches", default=5, type=int)
parser.add_argument("--cam_learning_rate", default=0.1, type=float)
parser.add_argument("--cam_weight_decay", default=1e-4, type=float)
parser.add_argument("--cam_eval_thres", default=0.15, type=float)
parser.add_argument("--cam_scales", default=(1.0, 0.5, 1.5, 2.0),
help="Multi-scale inferences")
parser.add_argument("--cam_weights_name", default="sess/res50_cam.pth", type=str)
parser.add_argument("--target_layer", default="stage4")
parser.add_argument("--cam_out_dir", default="result/cam_adv_mask", type=str)
parser.add_argument("--adv_iter", default=27, type=int)
parser.add_argument("--AD_coeff", default=7, type=int)
parser.add_argument("--AD_stepsize", default=0.08, type=float)
parser.add_argument("--score_th", default=0.5, type=float)
args = parser.parse_args()
torch.set_num_threads(1)
if not os.path.exists(args.cam_out_dir):
os.makedirs(args.cam_out_dir)
def adv_climb(image, epsilon, data_grad):
sign_data_grad = data_grad / (torch.max(torch.abs(data_grad))+1e-12)
perturbed_image = image + epsilon*sign_data_grad
perturbed_image = torch.clamp(perturbed_image, image.min().data.cpu().float(), image.max().data.cpu().float()) # min, max from data normalization
return perturbed_image
def add_discriminative(expanded_mask, regions, score_th):
region_ = regions / regions.max()
expanded_mask[region_>score_th]=1
return expanded_mask
def _work(process_id, model, dataset, args):
databin = dataset[process_id]
n_gpus = torch.cuda.device_count()
data_loader = DataLoader(databin, shuffle=False, num_workers=args.num_workers // n_gpus, pin_memory=True)
cam_sizes = [[], [], [], []] # scale 0,1,2,3
with cuda.device(process_id):
model.cuda()
gcam = GradCAM(model=model, candidate_layers=[args.target_layer]) # stage4
begin = time.time()
time_list = []
for iter, pack in enumerate(data_loader):
# print(pack['name']) # ['top_mosaic_09cm_area11_0_10']
img_name = pack['name'][0]
# if os.path.exists(os.path.join(args.cam_out_dir, img_name + '.npy')):
# continue
# if img_name != "top_mosaic_09cm_area1_4_3":
# continue
size = pack['size']
# print(size) # [tensor([256]), tensor([256])]
strided_size = imutils.get_strided_size(size, 4)
strided_up_size = imutils.get_strided_up_size(size, 16)
# print(strided_up_size)
outputs_cam = []
n_classes = len(list(torch.nonzero(pack['label'][0])[:, 0]))
if n_classes == 0:
print('类别为0')
continue
# print(n_classes)
begin1 = time.time()
for s_count, size_idx in enumerate([1, 0, 2, 3]):
orig_img = pack['img'][size_idx].clone()
for c_idx, c in enumerate(list(torch.nonzero(pack['label'][0])[:, 0])):
pack['img'][size_idx] = orig_img
# print(len(pack['img'])) # 4
# print(pack['img'][size_idx].shape) # torch.Size([1, 2, 3, 128, 128])
img_single = pack['img'][size_idx].detach()[0] # [:, 1]: flip
# 注意img_single有两个 一个是翻转前 一个是翻转后
# print('有类别，进来了:%s' % c)
if size_idx != 1:
total_adv_iter = args.adv_iter
else:
if args.adv_iter > 10:
total_adv_iter = args.adv_iter // 2 # 整除
mul_for_scale = 2
elif args.adv_iter < 6:
total_adv_iter = args.adv_iter
mul_for_scale = 1
else:
total_adv_iter = 5
mul_for_scale = float(total_adv_iter) / 5
for it in range(total_adv_iter):
img_single.requires_grad = True
# print(img_single.shape) # torch.Size([2, 3, 128, 128])
outputs = gcam.forward(img_single.cuda(non_blocking=True))
# print(outputs.shape) # torch.Size([2, 1, 8, 8])
if c_idx == 0 and it == 0:
cam_all_classes = torch.zeros([n_classes, outputs.shape[2], outputs.shape[3]])
# print(list(torch.nonzero(pack['label'][0])[:, 0]), c_idx, c)
# print(c.requires_grad)
gcam.backward(ids=c)
regions = gcam.generate(target_layer=args.target_layer)
# torch.Size([2, 1, 8, 8])
# print(list(torch.nonzero(pack['label'][0])[:, 0])) # [tensor(0)]
regions = regions[0] + regions[1].flip(-1)
# torch.Size([1, 8, 8])
##########生产伪标签过程中可视化背景和前景################
# fg_cam = regions.detach().cpu().numpy()[0]
# fg_img = np.uint8(255 * fg_cam)
# pseudo_img = cv2.applyColorMap(fg_img, cv2.COLORMAP_JET)
# pseudo_img = cv2.resize(pseudo_img,(256,256))
# cv2.imwrite('fg_grad_cam'+str(it)+'.png',pseudo_img)
# sys.exit(0)
#######################################################
if it == 0:
init_cam = regions.detach()
# print(cam_all_classes.shape, regions[0].shape)
cam_all_classes[c_idx] += regions[0].data.cpu() * mul_for_scale
# print(mul_for_scale) # 1
logit = outputs
logit = F.relu(logit)
# print(logit.shape) # torch.Size([2, 1, 8, 8])
logit = torchutils.gap2d(logit, keepdims=True)[:, :, 0, 0]
# print(logit.shape) # torch.Size([2, 1])
valid_cat = torch.nonzero(pack['label'][0])[:, 0]
logit_loss = - 2 * (logit[:, c]).sum() + torch.sum(logit)
expanded_mask = torch.zeros(regions.shape)
expanded_mask = add_discriminative(expanded_mask, regions, score_th=args.score_th)
L_AD = torch.sum((torch.abs(regions - init_cam))*expanded_mask.cuda())
loss = - logit_loss - L_AD * args.AD_coeff
model.zero_grad()
img_single.grad.zero_()
loss.backward()
data_grad = img_single.grad.data
perturbed_data = adv_climb(img_single, args.AD_stepsize, data_grad)
img_single = perturbed_data.detach()
outputs_cam.append(cam_all_classes)
strided_cam = torch.sum(torch.stack(
[F.interpolate(torch.unsqueeze(o, 0), strided_size, mode='bilinear', align_corners=False)[0] for o
in outputs_cam]), 0)
highres_cam = [F.interpolate(torch.unsqueeze(o, 1), strided_up_size,
mode='bilinear', align_corners=False) for o in outputs_cam]
# print(highres_cam[0].shape) # torch.Size([1, 1, 256, 256])
highres_cam = torch.sum(torch.stack(highres_cam, 0), 0)[:, 0, :size[0], :size[1]]
strided_cam /= F.adaptive_max_pool2d(strided_cam, (1, 1)) + 1e-5
highres_cam /= F.adaptive_max_pool2d(highres_cam, (1, 1)) + 1e-5
end1 = time.time()
time_list.append(end1-begin1)
print('图片' + img_name + '.png的CAM保存成功！！！', end1-begin1)
# if img_name == "top_mosaic_09cm_area1_4_3":
# sys.exit(0)
end = time.time()
print("total_time",(end-begin)/len(data_loader))
single_time = 0.0
for t in time_list:
single_time += t
print("single_total_time",single_time/len(time_list))
if __name__ == '__main__':
model = getattr(importlib.import_module(args.cam_network), 'CAM')()
model.load_state_dict(torch.load(args.cam_weights_name + '.pth'), strict=True)
model.eval()
n_gpus = torch.cuda.device_count()
print(n_gpus)
dataset = voc12.dataloader.VOC12ClassificationDatasetMSF(args.train_list,
voc12_root=args.voc12_root, scales=args.cam_scales)
dataset = torchutils.split_dataset(dataset, n_gpus)
# _work(0, model, dataset, args)
multiprocessing.spawn(_work, nprocs=n_gpus, args=(model, dataset, args), join=True)