├── .gitignore
├── .idea
├── encodings.xml
├── misc.xml
├── modules.xml
├── mymstn.iml
├── vcs.xml
└── workspace.xml
├── PretrainedAlexnet.py
├── README.md
├── data_list
├── amazon_list.txt
├── dslr_list.txt
└── webcam_list.txt
├── dataset.py
├── model.py
├── train.py
└── utils.py
/.gitignore:
--------------------------------------------------------------------------------
1 | *.pyc
2 | *.pth
3 | *.npy
4 | *.log
5 | *.pth
--------------------------------------------------------------------------------
/.idea/encodings.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
--------------------------------------------------------------------------------
/.idea/misc.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
--------------------------------------------------------------------------------
/.idea/modules.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
--------------------------------------------------------------------------------
/.idea/mymstn.iml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
11 |
--------------------------------------------------------------------------------
/.idea/vcs.xml:
--------------------------------------------------------------------------------
1 |
2 |
3 |
4 |
5 |
6 |
--------------------------------------------------------------------------------
/.idea/workspace.xml:
--------------------------------------------------------------------------------
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 |
82 |
83 |
84 |
85 | itertools
86 | Floss
87 | dreglos
88 | greglos
89 | n
90 | lr_mult_alex
91 | pred
92 | forward
93 | resume
94 | s_pred_label
95 | loss
96 | real
97 | sgd
98 | mse
99 | prealex
100 | bottle
101 | greg
102 | linearand
103 | data
104 | s_label
105 | v_label
106 | print
107 | optimizer
108 | mymodel
109 | vis
110 | conv
111 | args.name
112 | torchvision
113 | datasets
114 | n_class
115 |
116 |
117 | ../
118 | ../../
119 | self
120 | features
121 |
122 |
123 |
124 |
125 |
126 |
127 |
128 |
129 |
130 |
131 |
132 |
154 |
155 |
156 |
157 |
158 |
159 |
160 |
161 |
162 |
163 |
164 |
165 |
166 |
167 |
168 |
169 |
239 |
242 |
243 |
244 |
247 |
248 |
249 |
250 |
251 |
252 |
253 |
254 |
255 | 1545018637790
256 |
257 |
258 | 1545018637790
259 |
260 |
261 |
262 |
263 |
264 |
265 |
266 |
267 |
268 |
269 |
270 |
271 |
272 |
273 |
274 |
275 |
276 |
277 |
278 |
279 |
280 |
281 |
282 |
283 |
284 |
285 |
286 |
287 |
288 |
289 |
290 |
291 |
292 |
293 |
294 |
295 |
296 |
297 |
298 |
299 |
300 |
301 |
302 |
303 |
304 |
305 |
306 |
307 |
308 |
309 |
310 |
311 |
312 |
313 |
314 |
315 |
316 |
317 |
318 |
319 |
320 |
321 |
322 |
323 |
324 |
325 |
326 |
327 |
328 |
329 |
330 |
331 |
332 |
333 |
334 |
335 |
336 |
337 |
338 |
339 |
340 |
341 |
342 |
343 |
344 |
345 |
346 |
347 |
348 |
349 |
350 |
351 |
352 |
353 |
354 |
355 |
356 |
357 |
358 |
359 |
360 |
361 |
362 |
363 |
364 |
365 |
366 |
367 |
368 |
369 |
370 |
371 |
372 |
373 |
374 |
375 |
376 |
377 |
378 |
379 |
380 |
381 |
382 |
383 |
384 |
385 |
386 |
387 |
388 |
389 |
390 |
391 |
392 |
393 |
394 |
395 |
396 |
397 |
398 |
399 |
400 |
401 |
402 |
403 |
404 |
405 |
406 |
407 |
408 |
409 |
410 |
411 |
412 |
413 |
414 |
415 |
416 |
417 |
418 |
419 |
420 |
421 |
422 |
423 |
424 |
425 |
426 |
427 |
428 |
429 |
430 |
431 |
432 |
433 |
434 |
435 |
436 |
437 |
438 |
439 |
440 |
441 |
442 |
443 |
444 |
445 |
446 |
447 |
448 |
449 |
450 |
451 |
452 |
453 |
454 |
455 |
456 |
457 |
458 |
459 |
460 |
461 |
462 |
463 |
464 |
465 |
466 |
467 |
468 |
469 |
470 |
471 |
472 |
473 |
474 |
475 |
476 |
477 |
478 |
479 |
480 |
481 |
482 |
483 |
484 |
485 |
486 |
487 |
488 |
489 |
490 |
491 |
492 |
493 |
494 |
495 |
496 |
497 |
498 |
499 |
500 |
501 |
502 |
503 |
504 |
505 |
506 |
507 |
508 |
509 |
510 |
511 |
512 |
513 |
514 |
515 |
516 |
517 |
518 |
519 |
520 |
521 |
522 |
523 |
524 |
525 |
526 |
527 |
528 |
529 |
530 |
531 |
532 |
533 |
534 |
535 |
536 |
537 |
538 |
539 |
540 |
541 |
542 |
543 |
544 |
545 |
546 |
547 |
548 |
--------------------------------------------------------------------------------
/PretrainedAlexnet.py:
--------------------------------------------------------------------------------
1 | import itertools
2 | from itertools import chain
3 | import torch
4 | import torch.nn as nn
5 | import utils
6 | from torchvision.models import alexnet
7 |
8 | class AlexNet(nn.Module):
9 |
10 | def __init__(self, cudable, n_class):
11 | super(AlexNet, self).__init__()
12 | self.cudable = cudable
13 | self.n_class = n_class
14 | self.decay = 0.3
15 | self.s_centroid = torch.zeros(self.n_class, 256)
16 | self.t_centroid = torch.zeros(self.n_class, 256)
17 | if self.cudable:
18 | self.s_centroid = self.s_centroid.cuda()
19 | self.t_centroid = self.t_centroid.cuda()
20 | pretrained = alexnet(pretrained=True)
21 | self.features = pretrained.features
22 | self.classifier = nn.Sequential(*[pretrained.classifier[i] for i in range(6)])
23 | self.fc8 = nn.Sequential(
24 | nn.Linear(4096, 256)
25 | )
26 | self.fc9 = nn.Sequential(
27 | nn.Linear(256, self.n_class)
28 | )
29 | self.softmax = nn.Softmax(dim=0)
30 | self.D = nn.Sequential(
31 | nn.Linear(256, 1024),
32 | nn.ReLU(inplace=True),
33 | nn.Dropout(),
34 | nn.Linear(1024, 1024),
35 | nn.ReLU(inplace=True),
36 | nn.Dropout(),
37 | nn.Linear(1024, 1)
38 | )
39 | self.init()
40 |
41 | def init(self):
42 | self.init_linear(self.fc8[0])
43 | self.init_linear(self.fc9[0], std=0.005)
44 | self.init_linear(self.D[0],D=True)
45 | self.init_linear(self.D[3],D=True)
46 | self.init_linear(self.D[6],D=True, std=0.3)
47 | self.CEloss, self.MSEloss, self.BCEloss = nn.CrossEntropyLoss(), nn.MSELoss(), nn.BCEWithLogitsLoss(reduction='mean')
48 | if self.cudable:
49 | self.CEloss, self.MSEloss, self.BCEloss = self.CEloss.cuda(), self.MSEloss.cuda(), self.BCEloss.cuda()
50 |
51 | def init_linear(self, m, std=0.01, D=False):
52 | # nn.init.normal_(m.weight.data, 0, std)
53 | utils.truncated_normal_(m.weight.data, 0, std)
54 | # nn.init.xavier_normal_(m.weight)
55 | if D:
56 | m.bias.data.fill_(0)
57 | else:
58 | m.bias.data.fill_(0.1)
59 |
60 | def forward(self, x, training=True):
61 | conv_out = self.features(x)
62 | flattened = conv_out.view(conv_out.size(0), -1)
63 | dense_out = self.classifier(flattened)
64 | feature = self.fc8(dense_out)
65 | score = self.fc9(feature)
66 | pred = self.softmax(score)
67 | return feature, score, pred
68 |
69 | def forward_D(self, feature):
70 | logit = self.D(feature)
71 | return logit
72 |
73 | def closs(self, y_pred, y):
74 | C_loss = self.CEloss(y_pred, y)
75 | return C_loss
76 |
77 | def adloss(self, s_logits, t_logits, s_feature, t_feature, y_s, y_t):
78 | n, d = s_feature.shape
79 |
80 | # get labels
81 | s_labels, t_labels = y_s, torch.max(y_t, 1)[1]
82 |
83 | # image number in each class
84 | ones = torch.ones_like(s_labels, dtype=torch.float)
85 | zeros = torch.zeros(self.n_class)
86 | if self.cudable:
87 | zeros = zeros.cuda()
88 | s_n_classes = zeros.scatter_add(0, s_labels, ones)
89 | t_n_classes = zeros.scatter_add(0, t_labels, ones)
90 |
91 | # image number cannot be 0, when calculating centroids
92 | ones = torch.ones_like(s_n_classes)
93 | s_n_classes = torch.max(s_n_classes, ones)
94 | t_n_classes = torch.max(t_n_classes, ones)
95 |
96 | # calculating centroids, sum and divide
97 | zeros = torch.zeros(self.n_class, d)
98 | if self.cudable:
99 | zeros = zeros.cuda()
100 | s_sum_feature = zeros.scatter_add(0, torch.transpose(s_labels.repeat(d, 1), 1, 0), s_feature)
101 | t_sum_feature = zeros.scatter_add(0, torch.transpose(t_labels.repeat(d, 1), 1, 0), t_feature)
102 | current_s_centroid = torch.div(s_sum_feature, s_n_classes.view(self.n_class, 1))
103 | current_t_centroid = torch.div(t_sum_feature, t_n_classes.view(self.n_class, 1))
104 |
105 | # Moving Centroid
106 | decay = self.decay
107 | s_centroid = (1-decay) * self.s_centroid + decay * current_s_centroid
108 | t_centroid = (1-decay) * self.t_centroid + decay * current_t_centroid
109 | semantic_loss = self.MSEloss(s_centroid, t_centroid)
110 | self.s_centroid = s_centroid.detach()
111 | self.t_centroid = t_centroid.detach()
112 |
113 | # sigmoid binary cross entropy with reduce mean
114 | D_real_loss = self.BCEloss(t_logits, torch.ones_like(t_logits))
115 | D_fake_loss = self.BCEloss(s_logits, torch.zeros_like(s_logits))
116 | D_loss = (D_real_loss + D_fake_loss) * 0.1
117 | G_loss = -D_loss
118 |
119 | return G_loss, D_loss, semantic_loss
120 |
121 | # To some extent, can be replaced by weight_decay param of the optimizer.
122 | def regloss(self):
123 | Dregloss = [torch.sum(layer.weight ** 2) / 2 for layer in self.D if type(layer) == nn.Linear]
124 | layers = chain(self.features, self.classifier, self.fc8, self.fc9)
125 | Gregloss = [torch.sum(layer.weight ** 2) / 2 for layer in layers if type(layer) == nn.Conv2d or type(layer) == nn.Linear]
126 | mean = lambda x:0.0005 * torch.mean(torch.stack(x))
127 | return mean(Dregloss), mean(Gregloss)
128 |
129 |
130 | def get_optimizer(self, init_lr, lr_mult, lr_mult_D):
131 | w_finetune, b_finetune, w_train, b_train, w_D, b_D = [], [], [], [], [], []
132 |
133 | finetune_layers = itertools.chain(self.features, self.classifier)
134 | train_layers = itertools.chain(self.fc8, self.fc9)
135 | for layer in finetune_layers:
136 | if type(layer) == nn.Conv2d or type(layer) == nn.Linear:
137 | w_finetune.append(layer.weight)
138 | b_finetune.append(layer.bias)
139 | for layer in train_layers:
140 | if type(layer) == nn.Linear:
141 | w_train.append(layer.weight)
142 | b_train.append(layer.bias)
143 | for layer in self.D:
144 | if type(layer) == nn.Linear:
145 | w_D.append(layer.weight)
146 | b_D.append(layer.bias)
147 |
148 | opt = torch.optim.Adam([{'params': w_finetune, 'lr': init_lr * lr_mult[0]},
149 | {'params': b_finetune, 'lr': init_lr * lr_mult[1]},
150 | {'params': w_train, 'lr': init_lr * lr_mult[2]},
151 | {'params': b_train, 'lr': init_lr * lr_mult[3]}],
152 | lr=init_lr)
153 |
154 | opt_D = torch.optim.Adam([{'params': w_D, 'lr': init_lr * lr_mult_D[0]},
155 | {'params': b_D, 'lr': init_lr * lr_mult_D[1]}],
156 | lr=init_lr)
157 |
158 | return opt, opt_D
159 |
--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # PyTorch-MSTN
2 |
3 | pytorch reimplementation of [Moving Semantic Transfer Network](https://github.com/Mid-Push/Moving-Semantic-Transfer-Network)
4 |
5 | @inproceedings{xie2018learning,
6 | title={Learning Semantic Representations for Unsupervised Domain Adaptation},
7 | author={Xie, Shaoan and Zheng, Zibin and Chen, Liang and Chen, Chuan},
8 | booktitle={International Conference on Machine Learning},
9 | pages={5419--5428},
10 | year={2018}
11 | }
12 |
13 | ## Environment
14 |
15 | - Python 2.7
16 | - PyTorch 1.0.0
17 |
18 | ## Note
19 |
20 | - Amazon-Webcam实验复现成功，使用的是[pytorch_imagenet](https://github.com/jiecaoyu/pytorch_imagenet)中提供的pretrained weights和LRN实现。
21 | - MSTN作者使用的AlexNet来自于[Finetuning AlexNet with Tensorflow](https://github.com/kratzert/finetune_alexnet_with_tensorflow/)（和pytorch_imagenet的模型几乎相同）。我尝试[1]将这个模型及weights经过转换应用到PyTorch中；[2]使用torchvision提供的预训练的AlexNet（和[1]架构不同）。但这两种方式结果都只能到达67-70%。
22 | - 尝试了SGD和Adam，目前实验中momentum=0.9,init_lr=0.01的SGD的效果更好。
23 | - 对A-W任务，约在9000-12000次迭代时收敛。
24 | - 在将代码从TF迁移到PyTorch时可能需要注意的问题：[1]OpenCV默认的图像通道是BGR，而PyTorch（PIL）使用的通道一般是RGB（这个Repo没有转换通道）；[2]npy文件中的模型参数需要转置才能赋给PyTorch模型；[3]LRN层PyTorch已有官方实现，但它的参数size似乎和TF有所不同（这个Repo没有使用PyTorch的LRN层）；[4]PyTorch一般认为输入数据是0-1的，而caffe是0-255（这个Repo没有/255）。代码的迁移过程中还有很多没有理解的问题。
25 | - 在train.py中import model或PretrainedAlexnet来使用上面提到的[1][2]两种预训练AlexNet。如果要使用[2]的模型及weights，请在train.py中注释掉model.load_state_dict一行。
26 |
27 | ## Result
28 |
29 | | | Amazon-Webcam | Amazon-Dslr | Dslr-Webcam |
30 | | :--------------------: | :-----------: | :---------: | :---------: |
31 | | (paper)Source Only | 0.616 | 63.8 | 95.4 |
32 | | (paper)MSTN | 0.805 | 74.5 | 96.9 |
33 | | (this repo) Source Only| 0.618 | 63.2 | 95.6 |
34 | | (this repo) MSTN | 0.805 | 76.1 | 96.8 |
35 |
36 | ## Reference
37 |
38 | - [Moving Semantic Transfer Network](https://github.com/Mid-Push/Moving-Semantic-Transfer-Network)
39 | - [pytorch_imagenet](https://github.com/jiecaoyu/pytorch_imagenet)
40 | - [Finetuning AlexNet with Tensorflow](https://github.com/kratzert/finetune_alexnet_with_tensorflow/)
41 | - [Office31 dataset](https://people.eecs.berkeley.edu/~jhoffman/domainadapt/)
--------------------------------------------------------------------------------
/data_list/dslr_list.txt:
--------------------------------------------------------------------------------
1 | ../../dataset/office/dslr/images/calculator/frame_0001.jpg 5
2 | ../../dataset/office/dslr/images/calculator/frame_0002.jpg 5
3 | ../../dataset/office/dslr/images/calculator/frame_0003.jpg 5
4 | ../../dataset/office/dslr/images/calculator/frame_0004.jpg 5
5 | ../../dataset/office/dslr/images/calculator/frame_0005.jpg 5
6 | ../../dataset/office/dslr/images/calculator/frame_0006.jpg 5
7 | ../../dataset/office/dslr/images/calculator/frame_0007.jpg 5
8 | ../../dataset/office/dslr/images/calculator/frame_0008.jpg 5
9 | ../../dataset/office/dslr/images/calculator/frame_0009.jpg 5
10 | ../../dataset/office/dslr/images/calculator/frame_0010.jpg 5
11 | ../../dataset/office/dslr/images/calculator/frame_0011.jpg 5
12 | ../../dataset/office/dslr/images/calculator/frame_0012.jpg 5
13 | ../../dataset/office/dslr/images/ring_binder/frame_0001.jpg 24
14 | ../../dataset/office/dslr/images/ring_binder/frame_0002.jpg 24
15 | ../../dataset/office/dslr/images/ring_binder/frame_0003.jpg 24
16 | ../../dataset/office/dslr/images/ring_binder/frame_0004.jpg 24
17 | ../../dataset/office/dslr/images/ring_binder/frame_0005.jpg 24
18 | ../../dataset/office/dslr/images/ring_binder/frame_0006.jpg 24
19 | ../../dataset/office/dslr/images/ring_binder/frame_0007.jpg 24
20 | ../../dataset/office/dslr/images/ring_binder/frame_0008.jpg 24
21 | ../../dataset/office/dslr/images/ring_binder/frame_0009.jpg 24
22 | ../../dataset/office/dslr/images/ring_binder/frame_0010.jpg 24
23 | ../../dataset/office/dslr/images/printer/frame_0001.jpg 21
24 | ../../dataset/office/dslr/images/printer/frame_0002.jpg 21
25 | ../../dataset/office/dslr/images/printer/frame_0003.jpg 21
26 | ../../dataset/office/dslr/images/printer/frame_0004.jpg 21
27 | ../../dataset/office/dslr/images/printer/frame_0005.jpg 21
28 | ../../dataset/office/dslr/images/printer/frame_0006.jpg 21
29 | ../../dataset/office/dslr/images/printer/frame_0007.jpg 21
30 | ../../dataset/office/dslr/images/printer/frame_0008.jpg 21
31 | ../../dataset/office/dslr/images/printer/frame_0009.jpg 21
32 | ../../dataset/office/dslr/images/printer/frame_0010.jpg 21
33 | ../../dataset/office/dslr/images/printer/frame_0011.jpg 21
34 | ../../dataset/office/dslr/images/printer/frame_0012.jpg 21
35 | ../../dataset/office/dslr/images/printer/frame_0013.jpg 21
36 | ../../dataset/office/dslr/images/printer/frame_0014.jpg 21
37 | ../../dataset/office/dslr/images/printer/frame_0015.jpg 21
38 | ../../dataset/office/dslr/images/keyboard/frame_0001.jpg 11
39 | ../../dataset/office/dslr/images/keyboard/frame_0002.jpg 11
40 | ../../dataset/office/dslr/images/keyboard/frame_0003.jpg 11
41 | ../../dataset/office/dslr/images/keyboard/frame_0004.jpg 11
42 | ../../dataset/office/dslr/images/keyboard/frame_0005.jpg 11
43 | ../../dataset/office/dslr/images/keyboard/frame_0006.jpg 11
44 | ../../dataset/office/dslr/images/keyboard/frame_0007.jpg 11
45 | ../../dataset/office/dslr/images/keyboard/frame_0008.jpg 11
46 | ../../dataset/office/dslr/images/keyboard/frame_0009.jpg 11
47 | ../../dataset/office/dslr/images/keyboard/frame_0010.jpg 11
48 | ../../dataset/office/dslr/images/scissors/frame_0001.jpg 26
49 | ../../dataset/office/dslr/images/scissors/frame_0002.jpg 26
50 | ../../dataset/office/dslr/images/scissors/frame_0003.jpg 26
51 | ../../dataset/office/dslr/images/scissors/frame_0004.jpg 26
52 | ../../dataset/office/dslr/images/scissors/frame_0005.jpg 26
53 | ../../dataset/office/dslr/images/scissors/frame_0006.jpg 26
54 | ../../dataset/office/dslr/images/scissors/frame_0007.jpg 26
55 | ../../dataset/office/dslr/images/scissors/frame_0008.jpg 26
56 | ../../dataset/office/dslr/images/scissors/frame_0009.jpg 26
57 | ../../dataset/office/dslr/images/scissors/frame_0010.jpg 26
58 | ../../dataset/office/dslr/images/scissors/frame_0011.jpg 26
59 | ../../dataset/office/dslr/images/scissors/frame_0012.jpg 26
60 | ../../dataset/office/dslr/images/scissors/frame_0013.jpg 26
61 | ../../dataset/office/dslr/images/scissors/frame_0014.jpg 26
62 | ../../dataset/office/dslr/images/scissors/frame_0015.jpg 26
63 | ../../dataset/office/dslr/images/scissors/frame_0016.jpg 26
64 | ../../dataset/office/dslr/images/scissors/frame_0017.jpg 26
65 | ../../dataset/office/dslr/images/scissors/frame_0018.jpg 26
66 | ../../dataset/office/dslr/images/laptop_computer/frame_0001.jpg 12
67 | ../../dataset/office/dslr/images/laptop_computer/frame_0002.jpg 12
68 | ../../dataset/office/dslr/images/laptop_computer/frame_0003.jpg 12
69 | ../../dataset/office/dslr/images/laptop_computer/frame_0004.jpg 12
70 | ../../dataset/office/dslr/images/laptop_computer/frame_0005.jpg 12
71 | ../../dataset/office/dslr/images/laptop_computer/frame_0006.jpg 12
72 | ../../dataset/office/dslr/images/laptop_computer/frame_0007.jpg 12
73 | ../../dataset/office/dslr/images/laptop_computer/frame_0008.jpg 12
74 | ../../dataset/office/dslr/images/laptop_computer/frame_0009.jpg 12
75 | ../../dataset/office/dslr/images/laptop_computer/frame_0010.jpg 12
76 | ../../dataset/office/dslr/images/laptop_computer/frame_0011.jpg 12
77 | ../../dataset/office/dslr/images/laptop_computer/frame_0012.jpg 12
78 | ../../dataset/office/dslr/images/laptop_computer/frame_0013.jpg 12
79 | ../../dataset/office/dslr/images/laptop_computer/frame_0014.jpg 12
80 | ../../dataset/office/dslr/images/laptop_computer/frame_0015.jpg 12
81 | ../../dataset/office/dslr/images/laptop_computer/frame_0016.jpg 12
82 | ../../dataset/office/dslr/images/laptop_computer/frame_0017.jpg 12
83 | ../../dataset/office/dslr/images/laptop_computer/frame_0018.jpg 12
84 | ../../dataset/office/dslr/images/laptop_computer/frame_0019.jpg 12
85 | ../../dataset/office/dslr/images/laptop_computer/frame_0020.jpg 12
86 | ../../dataset/office/dslr/images/laptop_computer/frame_0021.jpg 12
87 | ../../dataset/office/dslr/images/laptop_computer/frame_0022.jpg 12
88 | ../../dataset/office/dslr/images/laptop_computer/frame_0023.jpg 12
89 | ../../dataset/office/dslr/images/laptop_computer/frame_0024.jpg 12
90 | ../../dataset/office/dslr/images/mouse/frame_0001.jpg 16
91 | ../../dataset/office/dslr/images/mouse/frame_0002.jpg 16
92 | ../../dataset/office/dslr/images/mouse/frame_0003.jpg 16
93 | ../../dataset/office/dslr/images/mouse/frame_0004.jpg 16
94 | ../../dataset/office/dslr/images/mouse/frame_0005.jpg 16
95 | ../../dataset/office/dslr/images/mouse/frame_0006.jpg 16
96 | ../../dataset/office/dslr/images/mouse/frame_0007.jpg 16
97 | ../../dataset/office/dslr/images/mouse/frame_0008.jpg 16
98 | ../../dataset/office/dslr/images/mouse/frame_0009.jpg 16
99 | ../../dataset/office/dslr/images/mouse/frame_0010.jpg 16
100 | ../../dataset/office/dslr/images/mouse/frame_0011.jpg 16
101 | ../../dataset/office/dslr/images/mouse/frame_0012.jpg 16
102 | ../../dataset/office/dslr/images/monitor/frame_0001.jpg 15
103 | ../../dataset/office/dslr/images/monitor/frame_0002.jpg 15
104 | ../../dataset/office/dslr/images/monitor/frame_0003.jpg 15
105 | ../../dataset/office/dslr/images/monitor/frame_0004.jpg 15
106 | ../../dataset/office/dslr/images/monitor/frame_0005.jpg 15
107 | ../../dataset/office/dslr/images/monitor/frame_0006.jpg 15
108 | ../../dataset/office/dslr/images/monitor/frame_0007.jpg 15
109 | ../../dataset/office/dslr/images/monitor/frame_0008.jpg 15
110 | ../../dataset/office/dslr/images/monitor/frame_0009.jpg 15
111 | ../../dataset/office/dslr/images/monitor/frame_0010.jpg 15
112 | ../../dataset/office/dslr/images/monitor/frame_0011.jpg 15
113 | ../../dataset/office/dslr/images/monitor/frame_0012.jpg 15
114 | ../../dataset/office/dslr/images/monitor/frame_0013.jpg 15
115 | ../../dataset/office/dslr/images/monitor/frame_0014.jpg 15
116 | ../../dataset/office/dslr/images/monitor/frame_0015.jpg 15
117 | ../../dataset/office/dslr/images/monitor/frame_0016.jpg 15
118 | ../../dataset/office/dslr/images/monitor/frame_0017.jpg 15
119 | ../../dataset/office/dslr/images/monitor/frame_0018.jpg 15
120 | ../../dataset/office/dslr/images/monitor/frame_0019.jpg 15
121 | ../../dataset/office/dslr/images/monitor/frame_0020.jpg 15
122 | ../../dataset/office/dslr/images/monitor/frame_0021.jpg 15
123 | ../../dataset/office/dslr/images/monitor/frame_0022.jpg 15
124 | ../../dataset/office/dslr/images/mug/frame_0001.jpg 17
125 | ../../dataset/office/dslr/images/mug/frame_0002.jpg 17
126 | ../../dataset/office/dslr/images/mug/frame_0003.jpg 17
127 | ../../dataset/office/dslr/images/mug/frame_0004.jpg 17
128 | ../../dataset/office/dslr/images/mug/frame_0005.jpg 17
129 | ../../dataset/office/dslr/images/mug/frame_0006.jpg 17
130 | ../../dataset/office/dslr/images/mug/frame_0007.jpg 17
131 | ../../dataset/office/dslr/images/mug/frame_0008.jpg 17
132 | ../../dataset/office/dslr/images/tape_dispenser/frame_0001.jpg 29
133 | ../../dataset/office/dslr/images/tape_dispenser/frame_0002.jpg 29
134 | ../../dataset/office/dslr/images/tape_dispenser/frame_0003.jpg 29
135 | ../../dataset/office/dslr/images/tape_dispenser/frame_0004.jpg 29
136 | ../../dataset/office/dslr/images/tape_dispenser/frame_0005.jpg 29
137 | ../../dataset/office/dslr/images/tape_dispenser/frame_0006.jpg 29
138 | ../../dataset/office/dslr/images/tape_dispenser/frame_0007.jpg 29
139 | ../../dataset/office/dslr/images/tape_dispenser/frame_0008.jpg 29
140 | ../../dataset/office/dslr/images/tape_dispenser/frame_0009.jpg 29
141 | ../../dataset/office/dslr/images/tape_dispenser/frame_0010.jpg 29
142 | ../../dataset/office/dslr/images/tape_dispenser/frame_0011.jpg 29
143 | ../../dataset/office/dslr/images/tape_dispenser/frame_0012.jpg 29
144 | ../../dataset/office/dslr/images/tape_dispenser/frame_0013.jpg 29
145 | ../../dataset/office/dslr/images/tape_dispenser/frame_0014.jpg 29
146 | ../../dataset/office/dslr/images/tape_dispenser/frame_0015.jpg 29
147 | ../../dataset/office/dslr/images/tape_dispenser/frame_0016.jpg 29
148 | ../../dataset/office/dslr/images/tape_dispenser/frame_0017.jpg 29
149 | ../../dataset/office/dslr/images/tape_dispenser/frame_0018.jpg 29
150 | ../../dataset/office/dslr/images/tape_dispenser/frame_0019.jpg 29
151 | ../../dataset/office/dslr/images/tape_dispenser/frame_0020.jpg 29
152 | ../../dataset/office/dslr/images/tape_dispenser/frame_0021.jpg 29
153 | ../../dataset/office/dslr/images/tape_dispenser/frame_0022.jpg 29
154 | ../../dataset/office/dslr/images/pen/frame_0001.jpg 19
155 | ../../dataset/office/dslr/images/pen/frame_0002.jpg 19
156 | ../../dataset/office/dslr/images/pen/frame_0003.jpg 19
157 | ../../dataset/office/dslr/images/pen/frame_0004.jpg 19
158 | ../../dataset/office/dslr/images/pen/frame_0005.jpg 19
159 | ../../dataset/office/dslr/images/pen/frame_0006.jpg 19
160 | ../../dataset/office/dslr/images/pen/frame_0007.jpg 19
161 | ../../dataset/office/dslr/images/pen/frame_0008.jpg 19
162 | ../../dataset/office/dslr/images/pen/frame_0009.jpg 19
163 | ../../dataset/office/dslr/images/pen/frame_0010.jpg 19
164 | ../../dataset/office/dslr/images/bike/frame_0001.jpg 1
165 | ../../dataset/office/dslr/images/bike/frame_0002.jpg 1
166 | ../../dataset/office/dslr/images/bike/frame_0003.jpg 1
167 | ../../dataset/office/dslr/images/bike/frame_0004.jpg 1
168 | ../../dataset/office/dslr/images/bike/frame_0005.jpg 1
169 | ../../dataset/office/dslr/images/bike/frame_0006.jpg 1
170 | ../../dataset/office/dslr/images/bike/frame_0007.jpg 1
171 | ../../dataset/office/dslr/images/bike/frame_0008.jpg 1
172 | ../../dataset/office/dslr/images/bike/frame_0009.jpg 1
173 | ../../dataset/office/dslr/images/bike/frame_0010.jpg 1
174 | ../../dataset/office/dslr/images/bike/frame_0011.jpg 1
175 | ../../dataset/office/dslr/images/bike/frame_0012.jpg 1
176 | ../../dataset/office/dslr/images/bike/frame_0013.jpg 1
177 | ../../dataset/office/dslr/images/bike/frame_0014.jpg 1
178 | ../../dataset/office/dslr/images/bike/frame_0015.jpg 1
179 | ../../dataset/office/dslr/images/bike/frame_0016.jpg 1
180 | ../../dataset/office/dslr/images/bike/frame_0017.jpg 1
181 | ../../dataset/office/dslr/images/bike/frame_0018.jpg 1
182 | ../../dataset/office/dslr/images/bike/frame_0019.jpg 1
183 | ../../dataset/office/dslr/images/bike/frame_0020.jpg 1
184 | ../../dataset/office/dslr/images/bike/frame_0021.jpg 1
185 | ../../dataset/office/dslr/images/punchers/frame_0001.jpg 23
186 | ../../dataset/office/dslr/images/punchers/frame_0002.jpg 23
187 | ../../dataset/office/dslr/images/punchers/frame_0003.jpg 23
188 | ../../dataset/office/dslr/images/punchers/frame_0004.jpg 23
189 | ../../dataset/office/dslr/images/punchers/frame_0005.jpg 23
190 | ../../dataset/office/dslr/images/punchers/frame_0006.jpg 23
191 | ../../dataset/office/dslr/images/punchers/frame_0007.jpg 23
192 | ../../dataset/office/dslr/images/punchers/frame_0008.jpg 23
193 | ../../dataset/office/dslr/images/punchers/frame_0009.jpg 23
194 | ../../dataset/office/dslr/images/punchers/frame_0010.jpg 23
195 | ../../dataset/office/dslr/images/punchers/frame_0011.jpg 23
196 | ../../dataset/office/dslr/images/punchers/frame_0012.jpg 23
197 | ../../dataset/office/dslr/images/punchers/frame_0013.jpg 23
198 | ../../dataset/office/dslr/images/punchers/frame_0014.jpg 23
199 | ../../dataset/office/dslr/images/punchers/frame_0015.jpg 23
200 | ../../dataset/office/dslr/images/punchers/frame_0016.jpg 23
201 | ../../dataset/office/dslr/images/punchers/frame_0017.jpg 23
202 | ../../dataset/office/dslr/images/punchers/frame_0018.jpg 23
203 | ../../dataset/office/dslr/images/back_pack/frame_0001.jpg 0
204 | ../../dataset/office/dslr/images/back_pack/frame_0002.jpg 0
205 | ../../dataset/office/dslr/images/back_pack/frame_0003.jpg 0
206 | ../../dataset/office/dslr/images/back_pack/frame_0004.jpg 0
207 | ../../dataset/office/dslr/images/back_pack/frame_0005.jpg 0
208 | ../../dataset/office/dslr/images/back_pack/frame_0006.jpg 0
209 | ../../dataset/office/dslr/images/back_pack/frame_0007.jpg 0
210 | ../../dataset/office/dslr/images/back_pack/frame_0008.jpg 0
211 | ../../dataset/office/dslr/images/back_pack/frame_0009.jpg 0
212 | ../../dataset/office/dslr/images/back_pack/frame_0010.jpg 0
213 | ../../dataset/office/dslr/images/back_pack/frame_0011.jpg 0
214 | ../../dataset/office/dslr/images/back_pack/frame_0012.jpg 0
215 | ../../dataset/office/dslr/images/desktop_computer/frame_0001.jpg 8
216 | ../../dataset/office/dslr/images/desktop_computer/frame_0002.jpg 8
217 | ../../dataset/office/dslr/images/desktop_computer/frame_0003.jpg 8
218 | ../../dataset/office/dslr/images/desktop_computer/frame_0004.jpg 8
219 | ../../dataset/office/dslr/images/desktop_computer/frame_0005.jpg 8
220 | ../../dataset/office/dslr/images/desktop_computer/frame_0006.jpg 8
221 | ../../dataset/office/dslr/images/desktop_computer/frame_0007.jpg 8
222 | ../../dataset/office/dslr/images/desktop_computer/frame_0008.jpg 8
223 | ../../dataset/office/dslr/images/desktop_computer/frame_0009.jpg 8
224 | ../../dataset/office/dslr/images/desktop_computer/frame_0010.jpg 8
225 | ../../dataset/office/dslr/images/desktop_computer/frame_0011.jpg 8
226 | ../../dataset/office/dslr/images/desktop_computer/frame_0012.jpg 8
227 | ../../dataset/office/dslr/images/desktop_computer/frame_0013.jpg 8
228 | ../../dataset/office/dslr/images/desktop_computer/frame_0014.jpg 8
229 | ../../dataset/office/dslr/images/desktop_computer/frame_0015.jpg 8
230 | ../../dataset/office/dslr/images/speaker/frame_0001.jpg 27
231 | ../../dataset/office/dslr/images/speaker/frame_0002.jpg 27
232 | ../../dataset/office/dslr/images/speaker/frame_0003.jpg 27
233 | ../../dataset/office/dslr/images/speaker/frame_0004.jpg 27
234 | ../../dataset/office/dslr/images/speaker/frame_0005.jpg 27
235 | ../../dataset/office/dslr/images/speaker/frame_0006.jpg 27
236 | ../../dataset/office/dslr/images/speaker/frame_0007.jpg 27
237 | ../../dataset/office/dslr/images/speaker/frame_0008.jpg 27
238 | ../../dataset/office/dslr/images/speaker/frame_0009.jpg 27
239 | ../../dataset/office/dslr/images/speaker/frame_0010.jpg 27
240 | ../../dataset/office/dslr/images/speaker/frame_0011.jpg 27
241 | ../../dataset/office/dslr/images/speaker/frame_0012.jpg 27
242 | ../../dataset/office/dslr/images/speaker/frame_0013.jpg 27
243 | ../../dataset/office/dslr/images/speaker/frame_0014.jpg 27
244 | ../../dataset/office/dslr/images/speaker/frame_0015.jpg 27
245 | ../../dataset/office/dslr/images/speaker/frame_0016.jpg 27
246 | ../../dataset/office/dslr/images/speaker/frame_0017.jpg 27
247 | ../../dataset/office/dslr/images/speaker/frame_0018.jpg 27
248 | ../../dataset/office/dslr/images/speaker/frame_0019.jpg 27
249 | ../../dataset/office/dslr/images/speaker/frame_0020.jpg 27
250 | ../../dataset/office/dslr/images/speaker/frame_0021.jpg 27
251 | ../../dataset/office/dslr/images/speaker/frame_0022.jpg 27
252 | ../../dataset/office/dslr/images/speaker/frame_0023.jpg 27
253 | ../../dataset/office/dslr/images/speaker/frame_0024.jpg 27
254 | ../../dataset/office/dslr/images/speaker/frame_0025.jpg 27
255 | ../../dataset/office/dslr/images/speaker/frame_0026.jpg 27
256 | ../../dataset/office/dslr/images/mobile_phone/frame_0001.jpg 14
257 | ../../dataset/office/dslr/images/mobile_phone/frame_0002.jpg 14
258 | ../../dataset/office/dslr/images/mobile_phone/frame_0003.jpg 14
259 | ../../dataset/office/dslr/images/mobile_phone/frame_0004.jpg 14
260 | ../../dataset/office/dslr/images/mobile_phone/frame_0005.jpg 14
261 | ../../dataset/office/dslr/images/mobile_phone/frame_0006.jpg 14
262 | ../../dataset/office/dslr/images/mobile_phone/frame_0007.jpg 14
263 | ../../dataset/office/dslr/images/mobile_phone/frame_0008.jpg 14
264 | ../../dataset/office/dslr/images/mobile_phone/frame_0009.jpg 14
265 | ../../dataset/office/dslr/images/mobile_phone/frame_0010.jpg 14
266 | ../../dataset/office/dslr/images/mobile_phone/frame_0011.jpg 14
267 | ../../dataset/office/dslr/images/mobile_phone/frame_0012.jpg 14
268 | ../../dataset/office/dslr/images/mobile_phone/frame_0013.jpg 14
269 | ../../dataset/office/dslr/images/mobile_phone/frame_0014.jpg 14
270 | ../../dataset/office/dslr/images/mobile_phone/frame_0015.jpg 14
271 | ../../dataset/office/dslr/images/mobile_phone/frame_0016.jpg 14
272 | ../../dataset/office/dslr/images/mobile_phone/frame_0017.jpg 14
273 | ../../dataset/office/dslr/images/mobile_phone/frame_0018.jpg 14
274 | ../../dataset/office/dslr/images/mobile_phone/frame_0019.jpg 14
275 | ../../dataset/office/dslr/images/mobile_phone/frame_0020.jpg 14
276 | ../../dataset/office/dslr/images/mobile_phone/frame_0021.jpg 14
277 | ../../dataset/office/dslr/images/mobile_phone/frame_0022.jpg 14
278 | ../../dataset/office/dslr/images/mobile_phone/frame_0023.jpg 14
279 | ../../dataset/office/dslr/images/mobile_phone/frame_0024.jpg 14
280 | ../../dataset/office/dslr/images/mobile_phone/frame_0025.jpg 14
281 | ../../dataset/office/dslr/images/mobile_phone/frame_0026.jpg 14
282 | ../../dataset/office/dslr/images/mobile_phone/frame_0027.jpg 14
283 | ../../dataset/office/dslr/images/mobile_phone/frame_0028.jpg 14
284 | ../../dataset/office/dslr/images/mobile_phone/frame_0029.jpg 14
285 | ../../dataset/office/dslr/images/mobile_phone/frame_0030.jpg 14
286 | ../../dataset/office/dslr/images/mobile_phone/frame_0031.jpg 14
287 | ../../dataset/office/dslr/images/paper_notebook/frame_0001.jpg 18
288 | ../../dataset/office/dslr/images/paper_notebook/frame_0002.jpg 18
289 | ../../dataset/office/dslr/images/paper_notebook/frame_0003.jpg 18
290 | ../../dataset/office/dslr/images/paper_notebook/frame_0004.jpg 18
291 | ../../dataset/office/dslr/images/paper_notebook/frame_0005.jpg 18
292 | ../../dataset/office/dslr/images/paper_notebook/frame_0006.jpg 18
293 | ../../dataset/office/dslr/images/paper_notebook/frame_0007.jpg 18
294 | ../../dataset/office/dslr/images/paper_notebook/frame_0008.jpg 18
295 | ../../dataset/office/dslr/images/paper_notebook/frame_0009.jpg 18
296 | ../../dataset/office/dslr/images/paper_notebook/frame_0010.jpg 18
297 | ../../dataset/office/dslr/images/ruler/frame_0001.jpg 25
298 | ../../dataset/office/dslr/images/ruler/frame_0002.jpg 25
299 | ../../dataset/office/dslr/images/ruler/frame_0003.jpg 25
300 | ../../dataset/office/dslr/images/ruler/frame_0004.jpg 25
301 | ../../dataset/office/dslr/images/ruler/frame_0005.jpg 25
302 | ../../dataset/office/dslr/images/ruler/frame_0006.jpg 25
303 | ../../dataset/office/dslr/images/ruler/frame_0007.jpg 25
304 | ../../dataset/office/dslr/images/letter_tray/frame_0001.jpg 13
305 | ../../dataset/office/dslr/images/letter_tray/frame_0002.jpg 13
306 | ../../dataset/office/dslr/images/letter_tray/frame_0003.jpg 13
307 | ../../dataset/office/dslr/images/letter_tray/frame_0004.jpg 13
308 | ../../dataset/office/dslr/images/letter_tray/frame_0005.jpg 13
309 | ../../dataset/office/dslr/images/letter_tray/frame_0006.jpg 13
310 | ../../dataset/office/dslr/images/letter_tray/frame_0007.jpg 13
311 | ../../dataset/office/dslr/images/letter_tray/frame_0008.jpg 13
312 | ../../dataset/office/dslr/images/letter_tray/frame_0009.jpg 13
313 | ../../dataset/office/dslr/images/letter_tray/frame_0010.jpg 13
314 | ../../dataset/office/dslr/images/letter_tray/frame_0011.jpg 13
315 | ../../dataset/office/dslr/images/letter_tray/frame_0012.jpg 13
316 | ../../dataset/office/dslr/images/letter_tray/frame_0013.jpg 13
317 | ../../dataset/office/dslr/images/letter_tray/frame_0014.jpg 13
318 | ../../dataset/office/dslr/images/letter_tray/frame_0015.jpg 13
319 | ../../dataset/office/dslr/images/letter_tray/frame_0016.jpg 13
320 | ../../dataset/office/dslr/images/file_cabinet/frame_0001.jpg 9
321 | ../../dataset/office/dslr/images/file_cabinet/frame_0002.jpg 9
322 | ../../dataset/office/dslr/images/file_cabinet/frame_0003.jpg 9
323 | ../../dataset/office/dslr/images/file_cabinet/frame_0004.jpg 9
324 | ../../dataset/office/dslr/images/file_cabinet/frame_0005.jpg 9
325 | ../../dataset/office/dslr/images/file_cabinet/frame_0006.jpg 9
326 | ../../dataset/office/dslr/images/file_cabinet/frame_0007.jpg 9
327 | ../../dataset/office/dslr/images/file_cabinet/frame_0008.jpg 9
328 | ../../dataset/office/dslr/images/file_cabinet/frame_0009.jpg 9
329 | ../../dataset/office/dslr/images/file_cabinet/frame_0010.jpg 9
330 | ../../dataset/office/dslr/images/file_cabinet/frame_0011.jpg 9
331 | ../../dataset/office/dslr/images/file_cabinet/frame_0012.jpg 9
332 | ../../dataset/office/dslr/images/file_cabinet/frame_0013.jpg 9
333 | ../../dataset/office/dslr/images/file_cabinet/frame_0014.jpg 9
334 | ../../dataset/office/dslr/images/file_cabinet/frame_0015.jpg 9
335 | ../../dataset/office/dslr/images/phone/frame_0001.jpg 20
336 | ../../dataset/office/dslr/images/phone/frame_0002.jpg 20
337 | ../../dataset/office/dslr/images/phone/frame_0003.jpg 20
338 | ../../dataset/office/dslr/images/phone/frame_0004.jpg 20
339 | ../../dataset/office/dslr/images/phone/frame_0005.jpg 20
340 | ../../dataset/office/dslr/images/phone/frame_0006.jpg 20
341 | ../../dataset/office/dslr/images/phone/frame_0007.jpg 20
342 | ../../dataset/office/dslr/images/phone/frame_0008.jpg 20
343 | ../../dataset/office/dslr/images/phone/frame_0009.jpg 20
344 | ../../dataset/office/dslr/images/phone/frame_0010.jpg 20
345 | ../../dataset/office/dslr/images/phone/frame_0011.jpg 20
346 | ../../dataset/office/dslr/images/phone/frame_0012.jpg 20
347 | ../../dataset/office/dslr/images/phone/frame_0013.jpg 20
348 | ../../dataset/office/dslr/images/bookcase/frame_0001.jpg 3
349 | ../../dataset/office/dslr/images/bookcase/frame_0002.jpg 3
350 | ../../dataset/office/dslr/images/bookcase/frame_0003.jpg 3
351 | ../../dataset/office/dslr/images/bookcase/frame_0004.jpg 3
352 | ../../dataset/office/dslr/images/bookcase/frame_0005.jpg 3
353 | ../../dataset/office/dslr/images/bookcase/frame_0006.jpg 3
354 | ../../dataset/office/dslr/images/bookcase/frame_0007.jpg 3
355 | ../../dataset/office/dslr/images/bookcase/frame_0008.jpg 3
356 | ../../dataset/office/dslr/images/bookcase/frame_0009.jpg 3
357 | ../../dataset/office/dslr/images/bookcase/frame_0010.jpg 3
358 | ../../dataset/office/dslr/images/bookcase/frame_0011.jpg 3
359 | ../../dataset/office/dslr/images/bookcase/frame_0012.jpg 3
360 | ../../dataset/office/dslr/images/projector/frame_0001.jpg 22
361 | ../../dataset/office/dslr/images/projector/frame_0002.jpg 22
362 | ../../dataset/office/dslr/images/projector/frame_0003.jpg 22
363 | ../../dataset/office/dslr/images/projector/frame_0004.jpg 22
364 | ../../dataset/office/dslr/images/projector/frame_0005.jpg 22
365 | ../../dataset/office/dslr/images/projector/frame_0006.jpg 22
366 | ../../dataset/office/dslr/images/projector/frame_0007.jpg 22
367 | ../../dataset/office/dslr/images/projector/frame_0008.jpg 22
368 | ../../dataset/office/dslr/images/projector/frame_0009.jpg 22
369 | ../../dataset/office/dslr/images/projector/frame_0010.jpg 22
370 | ../../dataset/office/dslr/images/projector/frame_0011.jpg 22
371 | ../../dataset/office/dslr/images/projector/frame_0012.jpg 22
372 | ../../dataset/office/dslr/images/projector/frame_0013.jpg 22
373 | ../../dataset/office/dslr/images/projector/frame_0014.jpg 22
374 | ../../dataset/office/dslr/images/projector/frame_0015.jpg 22
375 | ../../dataset/office/dslr/images/projector/frame_0016.jpg 22
376 | ../../dataset/office/dslr/images/projector/frame_0017.jpg 22
377 | ../../dataset/office/dslr/images/projector/frame_0018.jpg 22
378 | ../../dataset/office/dslr/images/projector/frame_0019.jpg 22
379 | ../../dataset/office/dslr/images/projector/frame_0020.jpg 22
380 | ../../dataset/office/dslr/images/projector/frame_0021.jpg 22
381 | ../../dataset/office/dslr/images/projector/frame_0022.jpg 22
382 | ../../dataset/office/dslr/images/projector/frame_0023.jpg 22
383 | ../../dataset/office/dslr/images/stapler/frame_0001.jpg 28
384 | ../../dataset/office/dslr/images/stapler/frame_0002.jpg 28
385 | ../../dataset/office/dslr/images/stapler/frame_0003.jpg 28
386 | ../../dataset/office/dslr/images/stapler/frame_0004.jpg 28
387 | ../../dataset/office/dslr/images/stapler/frame_0005.jpg 28
388 | ../../dataset/office/dslr/images/stapler/frame_0006.jpg 28
389 | ../../dataset/office/dslr/images/stapler/frame_0007.jpg 28
390 | ../../dataset/office/dslr/images/stapler/frame_0008.jpg 28
391 | ../../dataset/office/dslr/images/stapler/frame_0009.jpg 28
392 | ../../dataset/office/dslr/images/stapler/frame_0010.jpg 28
393 | ../../dataset/office/dslr/images/stapler/frame_0011.jpg 28
394 | ../../dataset/office/dslr/images/stapler/frame_0012.jpg 28
395 | ../../dataset/office/dslr/images/stapler/frame_0013.jpg 28
396 | ../../dataset/office/dslr/images/stapler/frame_0014.jpg 28
397 | ../../dataset/office/dslr/images/stapler/frame_0015.jpg 28
398 | ../../dataset/office/dslr/images/stapler/frame_0016.jpg 28
399 | ../../dataset/office/dslr/images/stapler/frame_0017.jpg 28
400 | ../../dataset/office/dslr/images/stapler/frame_0018.jpg 28
401 | ../../dataset/office/dslr/images/stapler/frame_0019.jpg 28
402 | ../../dataset/office/dslr/images/stapler/frame_0020.jpg 28
403 | ../../dataset/office/dslr/images/stapler/frame_0021.jpg 28
404 | ../../dataset/office/dslr/images/trash_can/frame_0001.jpg 30
405 | ../../dataset/office/dslr/images/trash_can/frame_0002.jpg 30
406 | ../../dataset/office/dslr/images/trash_can/frame_0003.jpg 30
407 | ../../dataset/office/dslr/images/trash_can/frame_0004.jpg 30
408 | ../../dataset/office/dslr/images/trash_can/frame_0005.jpg 30
409 | ../../dataset/office/dslr/images/trash_can/frame_0006.jpg 30
410 | ../../dataset/office/dslr/images/trash_can/frame_0007.jpg 30
411 | ../../dataset/office/dslr/images/trash_can/frame_0008.jpg 30
412 | ../../dataset/office/dslr/images/trash_can/frame_0009.jpg 30
413 | ../../dataset/office/dslr/images/trash_can/frame_0010.jpg 30
414 | ../../dataset/office/dslr/images/trash_can/frame_0011.jpg 30
415 | ../../dataset/office/dslr/images/trash_can/frame_0012.jpg 30
416 | ../../dataset/office/dslr/images/trash_can/frame_0013.jpg 30
417 | ../../dataset/office/dslr/images/trash_can/frame_0014.jpg 30
418 | ../../dataset/office/dslr/images/trash_can/frame_0015.jpg 30
419 | ../../dataset/office/dslr/images/bike_helmet/frame_0001.jpg 2
420 | ../../dataset/office/dslr/images/bike_helmet/frame_0002.jpg 2
421 | ../../dataset/office/dslr/images/bike_helmet/frame_0003.jpg 2
422 | ../../dataset/office/dslr/images/bike_helmet/frame_0004.jpg 2
423 | ../../dataset/office/dslr/images/bike_helmet/frame_0005.jpg 2
424 | ../../dataset/office/dslr/images/bike_helmet/frame_0006.jpg 2
425 | ../../dataset/office/dslr/images/bike_helmet/frame_0007.jpg 2
426 | ../../dataset/office/dslr/images/bike_helmet/frame_0008.jpg 2
427 | ../../dataset/office/dslr/images/bike_helmet/frame_0009.jpg 2
428 | ../../dataset/office/dslr/images/bike_helmet/frame_0010.jpg 2
429 | ../../dataset/office/dslr/images/bike_helmet/frame_0011.jpg 2
430 | ../../dataset/office/dslr/images/bike_helmet/frame_0012.jpg 2
431 | ../../dataset/office/dslr/images/bike_helmet/frame_0013.jpg 2
432 | ../../dataset/office/dslr/images/bike_helmet/frame_0014.jpg 2
433 | ../../dataset/office/dslr/images/bike_helmet/frame_0015.jpg 2
434 | ../../dataset/office/dslr/images/bike_helmet/frame_0016.jpg 2
435 | ../../dataset/office/dslr/images/bike_helmet/frame_0017.jpg 2
436 | ../../dataset/office/dslr/images/bike_helmet/frame_0018.jpg 2
437 | ../../dataset/office/dslr/images/bike_helmet/frame_0019.jpg 2
438 | ../../dataset/office/dslr/images/bike_helmet/frame_0020.jpg 2
439 | ../../dataset/office/dslr/images/bike_helmet/frame_0021.jpg 2
440 | ../../dataset/office/dslr/images/bike_helmet/frame_0022.jpg 2
441 | ../../dataset/office/dslr/images/bike_helmet/frame_0023.jpg 2
442 | ../../dataset/office/dslr/images/bike_helmet/frame_0024.jpg 2
443 | ../../dataset/office/dslr/images/headphones/frame_0001.jpg 10
444 | ../../dataset/office/dslr/images/headphones/frame_0002.jpg 10
445 | ../../dataset/office/dslr/images/headphones/frame_0003.jpg 10
446 | ../../dataset/office/dslr/images/headphones/frame_0004.jpg 10
447 | ../../dataset/office/dslr/images/headphones/frame_0005.jpg 10
448 | ../../dataset/office/dslr/images/headphones/frame_0006.jpg 10
449 | ../../dataset/office/dslr/images/headphones/frame_0007.jpg 10
450 | ../../dataset/office/dslr/images/headphones/frame_0008.jpg 10
451 | ../../dataset/office/dslr/images/headphones/frame_0009.jpg 10
452 | ../../dataset/office/dslr/images/headphones/frame_0010.jpg 10
453 | ../../dataset/office/dslr/images/headphones/frame_0011.jpg 10
454 | ../../dataset/office/dslr/images/headphones/frame_0012.jpg 10
455 | ../../dataset/office/dslr/images/headphones/frame_0013.jpg 10
456 | ../../dataset/office/dslr/images/desk_lamp/frame_0001.jpg 7
457 | ../../dataset/office/dslr/images/desk_lamp/frame_0002.jpg 7
458 | ../../dataset/office/dslr/images/desk_lamp/frame_0003.jpg 7
459 | ../../dataset/office/dslr/images/desk_lamp/frame_0004.jpg 7
460 | ../../dataset/office/dslr/images/desk_lamp/frame_0005.jpg 7
461 | ../../dataset/office/dslr/images/desk_lamp/frame_0006.jpg 7
462 | ../../dataset/office/dslr/images/desk_lamp/frame_0007.jpg 7
463 | ../../dataset/office/dslr/images/desk_lamp/frame_0008.jpg 7
464 | ../../dataset/office/dslr/images/desk_lamp/frame_0009.jpg 7
465 | ../../dataset/office/dslr/images/desk_lamp/frame_0010.jpg 7
466 | ../../dataset/office/dslr/images/desk_lamp/frame_0011.jpg 7
467 | ../../dataset/office/dslr/images/desk_lamp/frame_0012.jpg 7
468 | ../../dataset/office/dslr/images/desk_lamp/frame_0013.jpg 7
469 | ../../dataset/office/dslr/images/desk_lamp/frame_0014.jpg 7
470 | ../../dataset/office/dslr/images/desk_chair/frame_0001.jpg 6
471 | ../../dataset/office/dslr/images/desk_chair/frame_0002.jpg 6
472 | ../../dataset/office/dslr/images/desk_chair/frame_0003.jpg 6
473 | ../../dataset/office/dslr/images/desk_chair/frame_0004.jpg 6
474 | ../../dataset/office/dslr/images/desk_chair/frame_0005.jpg 6
475 | ../../dataset/office/dslr/images/desk_chair/frame_0006.jpg 6
476 | ../../dataset/office/dslr/images/desk_chair/frame_0007.jpg 6
477 | ../../dataset/office/dslr/images/desk_chair/frame_0008.jpg 6
478 | ../../dataset/office/dslr/images/desk_chair/frame_0009.jpg 6
479 | ../../dataset/office/dslr/images/desk_chair/frame_0010.jpg 6
480 | ../../dataset/office/dslr/images/desk_chair/frame_0011.jpg 6
481 | ../../dataset/office/dslr/images/desk_chair/frame_0012.jpg 6
482 | ../../dataset/office/dslr/images/desk_chair/frame_0013.jpg 6
483 | ../../dataset/office/dslr/images/bottle/frame_0001.jpg 4
484 | ../../dataset/office/dslr/images/bottle/frame_0002.jpg 4
485 | ../../dataset/office/dslr/images/bottle/frame_0003.jpg 4
486 | ../../dataset/office/dslr/images/bottle/frame_0004.jpg 4
487 | ../../dataset/office/dslr/images/bottle/frame_0005.jpg 4
488 | ../../dataset/office/dslr/images/bottle/frame_0006.jpg 4
489 | ../../dataset/office/dslr/images/bottle/frame_0007.jpg 4
490 | ../../dataset/office/dslr/images/bottle/frame_0008.jpg 4
491 | ../../dataset/office/dslr/images/bottle/frame_0009.jpg 4
492 | ../../dataset/office/dslr/images/bottle/frame_0010.jpg 4
493 | ../../dataset/office/dslr/images/bottle/frame_0011.jpg 4
494 | ../../dataset/office/dslr/images/bottle/frame_0012.jpg 4
495 | ../../dataset/office/dslr/images/bottle/frame_0013.jpg 4
496 | ../../dataset/office/dslr/images/bottle/frame_0014.jpg 4
497 | ../../dataset/office/dslr/images/bottle/frame_0015.jpg 4
498 | ../../dataset/office/dslr/images/bottle/frame_0016.jpg 4
499 |
--------------------------------------------------------------------------------
/data_list/webcam_list.txt:
--------------------------------------------------------------------------------
1 | ../../dataset/office/webcam/images/calculator/frame_0001.jpg 5
2 | ../../dataset/office/webcam/images/calculator/frame_0002.jpg 5
3 | ../../dataset/office/webcam/images/calculator/frame_0003.jpg 5
4 | ../../dataset/office/webcam/images/calculator/frame_0004.jpg 5
5 | ../../dataset/office/webcam/images/calculator/frame_0005.jpg 5
6 | ../../dataset/office/webcam/images/calculator/frame_0006.jpg 5
7 | ../../dataset/office/webcam/images/calculator/frame_0007.jpg 5
8 | ../../dataset/office/webcam/images/calculator/frame_0008.jpg 5
9 | ../../dataset/office/webcam/images/calculator/frame_0009.jpg 5
10 | ../../dataset/office/webcam/images/calculator/frame_0010.jpg 5
11 | ../../dataset/office/webcam/images/calculator/frame_0011.jpg 5
12 | ../../dataset/office/webcam/images/calculator/frame_0012.jpg 5
13 | ../../dataset/office/webcam/images/calculator/frame_0013.jpg 5
14 | ../../dataset/office/webcam/images/calculator/frame_0014.jpg 5
15 | ../../dataset/office/webcam/images/calculator/frame_0015.jpg 5
16 | ../../dataset/office/webcam/images/calculator/frame_0016.jpg 5
17 | ../../dataset/office/webcam/images/calculator/frame_0017.jpg 5
18 | ../../dataset/office/webcam/images/calculator/frame_0018.jpg 5
19 | ../../dataset/office/webcam/images/calculator/frame_0019.jpg 5
20 | ../../dataset/office/webcam/images/calculator/frame_0020.jpg 5
21 | ../../dataset/office/webcam/images/calculator/frame_0021.jpg 5
22 | ../../dataset/office/webcam/images/calculator/frame_0022.jpg 5
23 | ../../dataset/office/webcam/images/calculator/frame_0023.jpg 5
24 | ../../dataset/office/webcam/images/calculator/frame_0024.jpg 5
25 | ../../dataset/office/webcam/images/calculator/frame_0025.jpg 5
26 | ../../dataset/office/webcam/images/calculator/frame_0026.jpg 5
27 | ../../dataset/office/webcam/images/calculator/frame_0027.jpg 5
28 | ../../dataset/office/webcam/images/calculator/frame_0028.jpg 5
29 | ../../dataset/office/webcam/images/calculator/frame_0029.jpg 5
30 | ../../dataset/office/webcam/images/calculator/frame_0030.jpg 5
31 | ../../dataset/office/webcam/images/calculator/frame_0031.jpg 5
32 | ../../dataset/office/webcam/images/ring_binder/frame_0001.jpg 24
33 | ../../dataset/office/webcam/images/ring_binder/frame_0002.jpg 24
34 | ../../dataset/office/webcam/images/ring_binder/frame_0003.jpg 24
35 | ../../dataset/office/webcam/images/ring_binder/frame_0004.jpg 24
36 | ../../dataset/office/webcam/images/ring_binder/frame_0005.jpg 24
37 | ../../dataset/office/webcam/images/ring_binder/frame_0006.jpg 24
38 | ../../dataset/office/webcam/images/ring_binder/frame_0007.jpg 24
39 | ../../dataset/office/webcam/images/ring_binder/frame_0008.jpg 24
40 | ../../dataset/office/webcam/images/ring_binder/frame_0009.jpg 24
41 | ../../dataset/office/webcam/images/ring_binder/frame_0010.jpg 24
42 | ../../dataset/office/webcam/images/ring_binder/frame_0011.jpg 24
43 | ../../dataset/office/webcam/images/ring_binder/frame_0012.jpg 24
44 | ../../dataset/office/webcam/images/ring_binder/frame_0013.jpg 24
45 | ../../dataset/office/webcam/images/ring_binder/frame_0014.jpg 24
46 | ../../dataset/office/webcam/images/ring_binder/frame_0015.jpg 24
47 | ../../dataset/office/webcam/images/ring_binder/frame_0016.jpg 24
48 | ../../dataset/office/webcam/images/ring_binder/frame_0017.jpg 24
49 | ../../dataset/office/webcam/images/ring_binder/frame_0018.jpg 24
50 | ../../dataset/office/webcam/images/ring_binder/frame_0019.jpg 24
51 | ../../dataset/office/webcam/images/ring_binder/frame_0020.jpg 24
52 | ../../dataset/office/webcam/images/ring_binder/frame_0021.jpg 24
53 | ../../dataset/office/webcam/images/ring_binder/frame_0022.jpg 24
54 | ../../dataset/office/webcam/images/ring_binder/frame_0023.jpg 24
55 | ../../dataset/office/webcam/images/ring_binder/frame_0024.jpg 24
56 | ../../dataset/office/webcam/images/ring_binder/frame_0025.jpg 24
57 | ../../dataset/office/webcam/images/ring_binder/frame_0026.jpg 24
58 | ../../dataset/office/webcam/images/ring_binder/frame_0027.jpg 24
59 | ../../dataset/office/webcam/images/ring_binder/frame_0028.jpg 24
60 | ../../dataset/office/webcam/images/ring_binder/frame_0029.jpg 24
61 | ../../dataset/office/webcam/images/ring_binder/frame_0030.jpg 24
62 | ../../dataset/office/webcam/images/ring_binder/frame_0031.jpg 24
63 | ../../dataset/office/webcam/images/ring_binder/frame_0032.jpg 24
64 | ../../dataset/office/webcam/images/ring_binder/frame_0033.jpg 24
65 | ../../dataset/office/webcam/images/ring_binder/frame_0034.jpg 24
66 | ../../dataset/office/webcam/images/ring_binder/frame_0035.jpg 24
67 | ../../dataset/office/webcam/images/ring_binder/frame_0036.jpg 24
68 | ../../dataset/office/webcam/images/ring_binder/frame_0037.jpg 24
69 | ../../dataset/office/webcam/images/ring_binder/frame_0038.jpg 24
70 | ../../dataset/office/webcam/images/ring_binder/frame_0039.jpg 24
71 | ../../dataset/office/webcam/images/ring_binder/frame_0040.jpg 24
72 | ../../dataset/office/webcam/images/printer/frame_0001.jpg 21
73 | ../../dataset/office/webcam/images/printer/frame_0002.jpg 21
74 | ../../dataset/office/webcam/images/printer/frame_0003.jpg 21
75 | ../../dataset/office/webcam/images/printer/frame_0004.jpg 21
76 | ../../dataset/office/webcam/images/printer/frame_0005.jpg 21
77 | ../../dataset/office/webcam/images/printer/frame_0006.jpg 21
78 | ../../dataset/office/webcam/images/printer/frame_0007.jpg 21
79 | ../../dataset/office/webcam/images/printer/frame_0008.jpg 21
80 | ../../dataset/office/webcam/images/printer/frame_0009.jpg 21
81 | ../../dataset/office/webcam/images/printer/frame_0010.jpg 21
82 | ../../dataset/office/webcam/images/printer/frame_0011.jpg 21
83 | ../../dataset/office/webcam/images/printer/frame_0012.jpg 21
84 | ../../dataset/office/webcam/images/printer/frame_0013.jpg 21
85 | ../../dataset/office/webcam/images/printer/frame_0014.jpg 21
86 | ../../dataset/office/webcam/images/printer/frame_0015.jpg 21
87 | ../../dataset/office/webcam/images/printer/frame_0016.jpg 21
88 | ../../dataset/office/webcam/images/printer/frame_0017.jpg 21
89 | ../../dataset/office/webcam/images/printer/frame_0018.jpg 21
90 | ../../dataset/office/webcam/images/printer/frame_0019.jpg 21
91 | ../../dataset/office/webcam/images/printer/frame_0020.jpg 21
92 | ../../dataset/office/webcam/images/keyboard/frame_0001.jpg 11
93 | ../../dataset/office/webcam/images/keyboard/frame_0002.jpg 11
94 | ../../dataset/office/webcam/images/keyboard/frame_0003.jpg 11
95 | ../../dataset/office/webcam/images/keyboard/frame_0004.jpg 11
96 | ../../dataset/office/webcam/images/keyboard/frame_0005.jpg 11
97 | ../../dataset/office/webcam/images/keyboard/frame_0006.jpg 11
98 | ../../dataset/office/webcam/images/keyboard/frame_0007.jpg 11
99 | ../../dataset/office/webcam/images/keyboard/frame_0008.jpg 11
100 | ../../dataset/office/webcam/images/keyboard/frame_0009.jpg 11
101 | ../../dataset/office/webcam/images/keyboard/frame_0010.jpg 11
102 | ../../dataset/office/webcam/images/keyboard/frame_0011.jpg 11
103 | ../../dataset/office/webcam/images/keyboard/frame_0012.jpg 11
104 | ../../dataset/office/webcam/images/keyboard/frame_0013.jpg 11
105 | ../../dataset/office/webcam/images/keyboard/frame_0014.jpg 11
106 | ../../dataset/office/webcam/images/keyboard/frame_0015.jpg 11
107 | ../../dataset/office/webcam/images/keyboard/frame_0016.jpg 11
108 | ../../dataset/office/webcam/images/keyboard/frame_0017.jpg 11
109 | ../../dataset/office/webcam/images/keyboard/frame_0018.jpg 11
110 | ../../dataset/office/webcam/images/keyboard/frame_0019.jpg 11
111 | ../../dataset/office/webcam/images/keyboard/frame_0020.jpg 11
112 | ../../dataset/office/webcam/images/keyboard/frame_0021.jpg 11
113 | ../../dataset/office/webcam/images/keyboard/frame_0022.jpg 11
114 | ../../dataset/office/webcam/images/keyboard/frame_0023.jpg 11
115 | ../../dataset/office/webcam/images/keyboard/frame_0024.jpg 11
116 | ../../dataset/office/webcam/images/keyboard/frame_0025.jpg 11
117 | ../../dataset/office/webcam/images/keyboard/frame_0026.jpg 11
118 | ../../dataset/office/webcam/images/keyboard/frame_0027.jpg 11
119 | ../../dataset/office/webcam/images/scissors/frame_0001.jpg 26
120 | ../../dataset/office/webcam/images/scissors/frame_0002.jpg 26
121 | ../../dataset/office/webcam/images/scissors/frame_0003.jpg 26
122 | ../../dataset/office/webcam/images/scissors/frame_0004.jpg 26
123 | ../../dataset/office/webcam/images/scissors/frame_0005.jpg 26
124 | ../../dataset/office/webcam/images/scissors/frame_0006.jpg 26
125 | ../../dataset/office/webcam/images/scissors/frame_0007.jpg 26
126 | ../../dataset/office/webcam/images/scissors/frame_0008.jpg 26
127 | ../../dataset/office/webcam/images/scissors/frame_0009.jpg 26
128 | ../../dataset/office/webcam/images/scissors/frame_0010.jpg 26
129 | ../../dataset/office/webcam/images/scissors/frame_0011.jpg 26
130 | ../../dataset/office/webcam/images/scissors/frame_0012.jpg 26
131 | ../../dataset/office/webcam/images/scissors/frame_0013.jpg 26
132 | ../../dataset/office/webcam/images/scissors/frame_0014.jpg 26
133 | ../../dataset/office/webcam/images/scissors/frame_0015.jpg 26
134 | ../../dataset/office/webcam/images/scissors/frame_0016.jpg 26
135 | ../../dataset/office/webcam/images/scissors/frame_0017.jpg 26
136 | ../../dataset/office/webcam/images/scissors/frame_0018.jpg 26
137 | ../../dataset/office/webcam/images/scissors/frame_0019.jpg 26
138 | ../../dataset/office/webcam/images/scissors/frame_0020.jpg 26
139 | ../../dataset/office/webcam/images/scissors/frame_0021.jpg 26
140 | ../../dataset/office/webcam/images/scissors/frame_0022.jpg 26
141 | ../../dataset/office/webcam/images/scissors/frame_0023.jpg 26
142 | ../../dataset/office/webcam/images/scissors/frame_0024.jpg 26
143 | ../../dataset/office/webcam/images/scissors/frame_0025.jpg 26
144 | ../../dataset/office/webcam/images/laptop_computer/frame_0001.jpg 12
145 | ../../dataset/office/webcam/images/laptop_computer/frame_0002.jpg 12
146 | ../../dataset/office/webcam/images/laptop_computer/frame_0003.jpg 12
147 | ../../dataset/office/webcam/images/laptop_computer/frame_0004.jpg 12
148 | ../../dataset/office/webcam/images/laptop_computer/frame_0005.jpg 12
149 | ../../dataset/office/webcam/images/laptop_computer/frame_0006.jpg 12
150 | ../../dataset/office/webcam/images/laptop_computer/frame_0007.jpg 12
151 | ../../dataset/office/webcam/images/laptop_computer/frame_0008.jpg 12
152 | ../../dataset/office/webcam/images/laptop_computer/frame_0009.jpg 12
153 | ../../dataset/office/webcam/images/laptop_computer/frame_0010.jpg 12
154 | ../../dataset/office/webcam/images/laptop_computer/frame_0011.jpg 12
155 | ../../dataset/office/webcam/images/laptop_computer/frame_0012.jpg 12
156 | ../../dataset/office/webcam/images/laptop_computer/frame_0013.jpg 12
157 | ../../dataset/office/webcam/images/laptop_computer/frame_0014.jpg 12
158 | ../../dataset/office/webcam/images/laptop_computer/frame_0015.jpg 12
159 | ../../dataset/office/webcam/images/laptop_computer/frame_0016.jpg 12
160 | ../../dataset/office/webcam/images/laptop_computer/frame_0017.jpg 12
161 | ../../dataset/office/webcam/images/laptop_computer/frame_0018.jpg 12
162 | ../../dataset/office/webcam/images/laptop_computer/frame_0019.jpg 12
163 | ../../dataset/office/webcam/images/laptop_computer/frame_0020.jpg 12
164 | ../../dataset/office/webcam/images/laptop_computer/frame_0021.jpg 12
165 | ../../dataset/office/webcam/images/laptop_computer/frame_0022.jpg 12
166 | ../../dataset/office/webcam/images/laptop_computer/frame_0023.jpg 12
167 | ../../dataset/office/webcam/images/laptop_computer/frame_0024.jpg 12
168 | ../../dataset/office/webcam/images/laptop_computer/frame_0025.jpg 12
169 | ../../dataset/office/webcam/images/laptop_computer/frame_0026.jpg 12
170 | ../../dataset/office/webcam/images/laptop_computer/frame_0027.jpg 12
171 | ../../dataset/office/webcam/images/laptop_computer/frame_0028.jpg 12
172 | ../../dataset/office/webcam/images/laptop_computer/frame_0029.jpg 12
173 | ../../dataset/office/webcam/images/laptop_computer/frame_0030.jpg 12
174 | ../../dataset/office/webcam/images/mouse/frame_0001.jpg 16
175 | ../../dataset/office/webcam/images/mouse/frame_0002.jpg 16
176 | ../../dataset/office/webcam/images/mouse/frame_0003.jpg 16
177 | ../../dataset/office/webcam/images/mouse/frame_0004.jpg 16
178 | ../../dataset/office/webcam/images/mouse/frame_0005.jpg 16
179 | ../../dataset/office/webcam/images/mouse/frame_0006.jpg 16
180 | ../../dataset/office/webcam/images/mouse/frame_0007.jpg 16
181 | ../../dataset/office/webcam/images/mouse/frame_0008.jpg 16
182 | ../../dataset/office/webcam/images/mouse/frame_0009.jpg 16
183 | ../../dataset/office/webcam/images/mouse/frame_0010.jpg 16
184 | ../../dataset/office/webcam/images/mouse/frame_0011.jpg 16
185 | ../../dataset/office/webcam/images/mouse/frame_0012.jpg 16
186 | ../../dataset/office/webcam/images/mouse/frame_0013.jpg 16
187 | ../../dataset/office/webcam/images/mouse/frame_0014.jpg 16
188 | ../../dataset/office/webcam/images/mouse/frame_0015.jpg 16
189 | ../../dataset/office/webcam/images/mouse/frame_0016.jpg 16
190 | ../../dataset/office/webcam/images/mouse/frame_0017.jpg 16
191 | ../../dataset/office/webcam/images/mouse/frame_0018.jpg 16
192 | ../../dataset/office/webcam/images/mouse/frame_0019.jpg 16
193 | ../../dataset/office/webcam/images/mouse/frame_0020.jpg 16
194 | ../../dataset/office/webcam/images/mouse/frame_0021.jpg 16
195 | ../../dataset/office/webcam/images/mouse/frame_0022.jpg 16
196 | ../../dataset/office/webcam/images/mouse/frame_0023.jpg 16
197 | ../../dataset/office/webcam/images/mouse/frame_0024.jpg 16
198 | ../../dataset/office/webcam/images/mouse/frame_0025.jpg 16
199 | ../../dataset/office/webcam/images/mouse/frame_0026.jpg 16
200 | ../../dataset/office/webcam/images/mouse/frame_0027.jpg 16
201 | ../../dataset/office/webcam/images/mouse/frame_0028.jpg 16
202 | ../../dataset/office/webcam/images/mouse/frame_0029.jpg 16
203 | ../../dataset/office/webcam/images/mouse/frame_0030.jpg 16
204 | ../../dataset/office/webcam/images/monitor/frame_0001.jpg 15
205 | ../../dataset/office/webcam/images/monitor/frame_0002.jpg 15
206 | ../../dataset/office/webcam/images/monitor/frame_0003.jpg 15
207 | ../../dataset/office/webcam/images/monitor/frame_0004.jpg 15
208 | ../../dataset/office/webcam/images/monitor/frame_0005.jpg 15
209 | ../../dataset/office/webcam/images/monitor/frame_0006.jpg 15
210 | ../../dataset/office/webcam/images/monitor/frame_0007.jpg 15
211 | ../../dataset/office/webcam/images/monitor/frame_0008.jpg 15
212 | ../../dataset/office/webcam/images/monitor/frame_0009.jpg 15
213 | ../../dataset/office/webcam/images/monitor/frame_0010.jpg 15
214 | ../../dataset/office/webcam/images/monitor/frame_0011.jpg 15
215 | ../../dataset/office/webcam/images/monitor/frame_0012.jpg 15
216 | ../../dataset/office/webcam/images/monitor/frame_0013.jpg 15
217 | ../../dataset/office/webcam/images/monitor/frame_0014.jpg 15
218 | ../../dataset/office/webcam/images/monitor/frame_0015.jpg 15
219 | ../../dataset/office/webcam/images/monitor/frame_0016.jpg 15
220 | ../../dataset/office/webcam/images/monitor/frame_0017.jpg 15
221 | ../../dataset/office/webcam/images/monitor/frame_0018.jpg 15
222 | ../../dataset/office/webcam/images/monitor/frame_0019.jpg 15
223 | ../../dataset/office/webcam/images/monitor/frame_0020.jpg 15
224 | ../../dataset/office/webcam/images/monitor/frame_0021.jpg 15
225 | ../../dataset/office/webcam/images/monitor/frame_0022.jpg 15
226 | ../../dataset/office/webcam/images/monitor/frame_0023.jpg 15
227 | ../../dataset/office/webcam/images/monitor/frame_0024.jpg 15
228 | ../../dataset/office/webcam/images/monitor/frame_0025.jpg 15
229 | ../../dataset/office/webcam/images/monitor/frame_0026.jpg 15
230 | ../../dataset/office/webcam/images/monitor/frame_0027.jpg 15
231 | ../../dataset/office/webcam/images/monitor/frame_0028.jpg 15
232 | ../../dataset/office/webcam/images/monitor/frame_0029.jpg 15
233 | ../../dataset/office/webcam/images/monitor/frame_0030.jpg 15
234 | ../../dataset/office/webcam/images/monitor/frame_0031.jpg 15
235 | ../../dataset/office/webcam/images/monitor/frame_0032.jpg 15
236 | ../../dataset/office/webcam/images/monitor/frame_0033.jpg 15
237 | ../../dataset/office/webcam/images/monitor/frame_0034.jpg 15
238 | ../../dataset/office/webcam/images/monitor/frame_0035.jpg 15
239 | ../../dataset/office/webcam/images/monitor/frame_0036.jpg 15
240 | ../../dataset/office/webcam/images/monitor/frame_0037.jpg 15
241 | ../../dataset/office/webcam/images/monitor/frame_0038.jpg 15
242 | ../../dataset/office/webcam/images/monitor/frame_0039.jpg 15
243 | ../../dataset/office/webcam/images/monitor/frame_0040.jpg 15
244 | ../../dataset/office/webcam/images/monitor/frame_0041.jpg 15
245 | ../../dataset/office/webcam/images/monitor/frame_0042.jpg 15
246 | ../../dataset/office/webcam/images/monitor/frame_0043.jpg 15
247 | ../../dataset/office/webcam/images/mug/frame_0001.jpg 17
248 | ../../dataset/office/webcam/images/mug/frame_0002.jpg 17
249 | ../../dataset/office/webcam/images/mug/frame_0003.jpg 17
250 | ../../dataset/office/webcam/images/mug/frame_0004.jpg 17
251 | ../../dataset/office/webcam/images/mug/frame_0005.jpg 17
252 | ../../dataset/office/webcam/images/mug/frame_0006.jpg 17
253 | ../../dataset/office/webcam/images/mug/frame_0007.jpg 17
254 | ../../dataset/office/webcam/images/mug/frame_0008.jpg 17
255 | ../../dataset/office/webcam/images/mug/frame_0009.jpg 17
256 | ../../dataset/office/webcam/images/mug/frame_0010.jpg 17
257 | ../../dataset/office/webcam/images/mug/frame_0011.jpg 17
258 | ../../dataset/office/webcam/images/mug/frame_0012.jpg 17
259 | ../../dataset/office/webcam/images/mug/frame_0013.jpg 17
260 | ../../dataset/office/webcam/images/mug/frame_0014.jpg 17
261 | ../../dataset/office/webcam/images/mug/frame_0015.jpg 17
262 | ../../dataset/office/webcam/images/mug/frame_0016.jpg 17
263 | ../../dataset/office/webcam/images/mug/frame_0017.jpg 17
264 | ../../dataset/office/webcam/images/mug/frame_0018.jpg 17
265 | ../../dataset/office/webcam/images/mug/frame_0019.jpg 17
266 | ../../dataset/office/webcam/images/mug/frame_0020.jpg 17
267 | ../../dataset/office/webcam/images/mug/frame_0021.jpg 17
268 | ../../dataset/office/webcam/images/mug/frame_0022.jpg 17
269 | ../../dataset/office/webcam/images/mug/frame_0023.jpg 17
270 | ../../dataset/office/webcam/images/mug/frame_0024.jpg 17
271 | ../../dataset/office/webcam/images/mug/frame_0025.jpg 17
272 | ../../dataset/office/webcam/images/mug/frame_0026.jpg 17
273 | ../../dataset/office/webcam/images/mug/frame_0027.jpg 17
274 | ../../dataset/office/webcam/images/tape_dispenser/frame_0001.jpg 29
275 | ../../dataset/office/webcam/images/tape_dispenser/frame_0002.jpg 29
276 | ../../dataset/office/webcam/images/tape_dispenser/frame_0003.jpg 29
277 | ../../dataset/office/webcam/images/tape_dispenser/frame_0004.jpg 29
278 | ../../dataset/office/webcam/images/tape_dispenser/frame_0005.jpg 29
279 | ../../dataset/office/webcam/images/tape_dispenser/frame_0006.jpg 29
280 | ../../dataset/office/webcam/images/tape_dispenser/frame_0007.jpg 29
281 | ../../dataset/office/webcam/images/tape_dispenser/frame_0008.jpg 29
282 | ../../dataset/office/webcam/images/tape_dispenser/frame_0009.jpg 29
283 | ../../dataset/office/webcam/images/tape_dispenser/frame_0010.jpg 29
284 | ../../dataset/office/webcam/images/tape_dispenser/frame_0011.jpg 29
285 | ../../dataset/office/webcam/images/tape_dispenser/frame_0012.jpg 29
286 | ../../dataset/office/webcam/images/tape_dispenser/frame_0013.jpg 29
287 | ../../dataset/office/webcam/images/tape_dispenser/frame_0014.jpg 29
288 | ../../dataset/office/webcam/images/tape_dispenser/frame_0015.jpg 29
289 | ../../dataset/office/webcam/images/tape_dispenser/frame_0016.jpg 29
290 | ../../dataset/office/webcam/images/tape_dispenser/frame_0017.jpg 29
291 | ../../dataset/office/webcam/images/tape_dispenser/frame_0018.jpg 29
292 | ../../dataset/office/webcam/images/tape_dispenser/frame_0019.jpg 29
293 | ../../dataset/office/webcam/images/tape_dispenser/frame_0020.jpg 29
294 | ../../dataset/office/webcam/images/tape_dispenser/frame_0021.jpg 29
295 | ../../dataset/office/webcam/images/tape_dispenser/frame_0022.jpg 29
296 | ../../dataset/office/webcam/images/tape_dispenser/frame_0023.jpg 29
297 | ../../dataset/office/webcam/images/pen/frame_0001.jpg 19
298 | ../../dataset/office/webcam/images/pen/frame_0002.jpg 19
299 | ../../dataset/office/webcam/images/pen/frame_0003.jpg 19
300 | ../../dataset/office/webcam/images/pen/frame_0004.jpg 19
301 | ../../dataset/office/webcam/images/pen/frame_0005.jpg 19
302 | ../../dataset/office/webcam/images/pen/frame_0006.jpg 19
303 | ../../dataset/office/webcam/images/pen/frame_0007.jpg 19
304 | ../../dataset/office/webcam/images/pen/frame_0008.jpg 19
305 | ../../dataset/office/webcam/images/pen/frame_0009.jpg 19
306 | ../../dataset/office/webcam/images/pen/frame_0010.jpg 19
307 | ../../dataset/office/webcam/images/pen/frame_0011.jpg 19
308 | ../../dataset/office/webcam/images/pen/frame_0012.jpg 19
309 | ../../dataset/office/webcam/images/pen/frame_0013.jpg 19
310 | ../../dataset/office/webcam/images/pen/frame_0014.jpg 19
311 | ../../dataset/office/webcam/images/pen/frame_0015.jpg 19
312 | ../../dataset/office/webcam/images/pen/frame_0016.jpg 19
313 | ../../dataset/office/webcam/images/pen/frame_0017.jpg 19
314 | ../../dataset/office/webcam/images/pen/frame_0018.jpg 19
315 | ../../dataset/office/webcam/images/pen/frame_0019.jpg 19
316 | ../../dataset/office/webcam/images/pen/frame_0020.jpg 19
317 | ../../dataset/office/webcam/images/pen/frame_0021.jpg 19
318 | ../../dataset/office/webcam/images/pen/frame_0022.jpg 19
319 | ../../dataset/office/webcam/images/pen/frame_0023.jpg 19
320 | ../../dataset/office/webcam/images/pen/frame_0024.jpg 19
321 | ../../dataset/office/webcam/images/pen/frame_0025.jpg 19
322 | ../../dataset/office/webcam/images/pen/frame_0026.jpg 19
323 | ../../dataset/office/webcam/images/pen/frame_0027.jpg 19
324 | ../../dataset/office/webcam/images/pen/frame_0028.jpg 19
325 | ../../dataset/office/webcam/images/pen/frame_0029.jpg 19
326 | ../../dataset/office/webcam/images/pen/frame_0030.jpg 19
327 | ../../dataset/office/webcam/images/pen/frame_0031.jpg 19
328 | ../../dataset/office/webcam/images/pen/frame_0032.jpg 19
329 | ../../dataset/office/webcam/images/bike/frame_0001.jpg 1
330 | ../../dataset/office/webcam/images/bike/frame_0002.jpg 1
331 | ../../dataset/office/webcam/images/bike/frame_0003.jpg 1
332 | ../../dataset/office/webcam/images/bike/frame_0004.jpg 1
333 | ../../dataset/office/webcam/images/bike/frame_0005.jpg 1
334 | ../../dataset/office/webcam/images/bike/frame_0006.jpg 1
335 | ../../dataset/office/webcam/images/bike/frame_0007.jpg 1
336 | ../../dataset/office/webcam/images/bike/frame_0008.jpg 1
337 | ../../dataset/office/webcam/images/bike/frame_0009.jpg 1
338 | ../../dataset/office/webcam/images/bike/frame_0010.jpg 1
339 | ../../dataset/office/webcam/images/bike/frame_0011.jpg 1
340 | ../../dataset/office/webcam/images/bike/frame_0012.jpg 1
341 | ../../dataset/office/webcam/images/bike/frame_0013.jpg 1
342 | ../../dataset/office/webcam/images/bike/frame_0014.jpg 1
343 | ../../dataset/office/webcam/images/bike/frame_0015.jpg 1
344 | ../../dataset/office/webcam/images/bike/frame_0016.jpg 1
345 | ../../dataset/office/webcam/images/bike/frame_0017.jpg 1
346 | ../../dataset/office/webcam/images/bike/frame_0018.jpg 1
347 | ../../dataset/office/webcam/images/bike/frame_0019.jpg 1
348 | ../../dataset/office/webcam/images/bike/frame_0020.jpg 1
349 | ../../dataset/office/webcam/images/bike/frame_0021.jpg 1
350 | ../../dataset/office/webcam/images/punchers/frame_0001.jpg 23
351 | ../../dataset/office/webcam/images/punchers/frame_0002.jpg 23
352 | ../../dataset/office/webcam/images/punchers/frame_0003.jpg 23
353 | ../../dataset/office/webcam/images/punchers/frame_0004.jpg 23
354 | ../../dataset/office/webcam/images/punchers/frame_0005.jpg 23
355 | ../../dataset/office/webcam/images/punchers/frame_0006.jpg 23
356 | ../../dataset/office/webcam/images/punchers/frame_0007.jpg 23
357 | ../../dataset/office/webcam/images/punchers/frame_0008.jpg 23
358 | ../../dataset/office/webcam/images/punchers/frame_0009.jpg 23
359 | ../../dataset/office/webcam/images/punchers/frame_0010.jpg 23
360 | ../../dataset/office/webcam/images/punchers/frame_0011.jpg 23
361 | ../../dataset/office/webcam/images/punchers/frame_0012.jpg 23
362 | ../../dataset/office/webcam/images/punchers/frame_0013.jpg 23
363 | ../../dataset/office/webcam/images/punchers/frame_0014.jpg 23
364 | ../../dataset/office/webcam/images/punchers/frame_0015.jpg 23
365 | ../../dataset/office/webcam/images/punchers/frame_0016.jpg 23
366 | ../../dataset/office/webcam/images/punchers/frame_0017.jpg 23
367 | ../../dataset/office/webcam/images/punchers/frame_0018.jpg 23
368 | ../../dataset/office/webcam/images/punchers/frame_0019.jpg 23
369 | ../../dataset/office/webcam/images/punchers/frame_0020.jpg 23
370 | ../../dataset/office/webcam/images/punchers/frame_0021.jpg 23
371 | ../../dataset/office/webcam/images/punchers/frame_0022.jpg 23
372 | ../../dataset/office/webcam/images/punchers/frame_0023.jpg 23
373 | ../../dataset/office/webcam/images/punchers/frame_0024.jpg 23
374 | ../../dataset/office/webcam/images/punchers/frame_0025.jpg 23
375 | ../../dataset/office/webcam/images/punchers/frame_0026.jpg 23
376 | ../../dataset/office/webcam/images/punchers/frame_0027.jpg 23
377 | ../../dataset/office/webcam/images/back_pack/frame_0001.jpg 0
378 | ../../dataset/office/webcam/images/back_pack/frame_0002.jpg 0
379 | ../../dataset/office/webcam/images/back_pack/frame_0003.jpg 0
380 | ../../dataset/office/webcam/images/back_pack/frame_0004.jpg 0
381 | ../../dataset/office/webcam/images/back_pack/frame_0005.jpg 0
382 | ../../dataset/office/webcam/images/back_pack/frame_0006.jpg 0
383 | ../../dataset/office/webcam/images/back_pack/frame_0007.jpg 0
384 | ../../dataset/office/webcam/images/back_pack/frame_0008.jpg 0
385 | ../../dataset/office/webcam/images/back_pack/frame_0009.jpg 0
386 | ../../dataset/office/webcam/images/back_pack/frame_0010.jpg 0
387 | ../../dataset/office/webcam/images/back_pack/frame_0011.jpg 0
388 | ../../dataset/office/webcam/images/back_pack/frame_0012.jpg 0
389 | ../../dataset/office/webcam/images/back_pack/frame_0013.jpg 0
390 | ../../dataset/office/webcam/images/back_pack/frame_0014.jpg 0
391 | ../../dataset/office/webcam/images/back_pack/frame_0015.jpg 0
392 | ../../dataset/office/webcam/images/back_pack/frame_0016.jpg 0
393 | ../../dataset/office/webcam/images/back_pack/frame_0017.jpg 0
394 | ../../dataset/office/webcam/images/back_pack/frame_0018.jpg 0
395 | ../../dataset/office/webcam/images/back_pack/frame_0019.jpg 0
396 | ../../dataset/office/webcam/images/back_pack/frame_0020.jpg 0
397 | ../../dataset/office/webcam/images/back_pack/frame_0021.jpg 0
398 | ../../dataset/office/webcam/images/back_pack/frame_0022.jpg 0
399 | ../../dataset/office/webcam/images/back_pack/frame_0023.jpg 0
400 | ../../dataset/office/webcam/images/back_pack/frame_0024.jpg 0
401 | ../../dataset/office/webcam/images/back_pack/frame_0025.jpg 0
402 | ../../dataset/office/webcam/images/back_pack/frame_0026.jpg 0
403 | ../../dataset/office/webcam/images/back_pack/frame_0027.jpg 0
404 | ../../dataset/office/webcam/images/back_pack/frame_0028.jpg 0
405 | ../../dataset/office/webcam/images/back_pack/frame_0029.jpg 0
406 | ../../dataset/office/webcam/images/desktop_computer/frame_0001.jpg 8
407 | ../../dataset/office/webcam/images/desktop_computer/frame_0002.jpg 8
408 | ../../dataset/office/webcam/images/desktop_computer/frame_0003.jpg 8
409 | ../../dataset/office/webcam/images/desktop_computer/frame_0004.jpg 8
410 | ../../dataset/office/webcam/images/desktop_computer/frame_0005.jpg 8
411 | ../../dataset/office/webcam/images/desktop_computer/frame_0006.jpg 8
412 | ../../dataset/office/webcam/images/desktop_computer/frame_0007.jpg 8
413 | ../../dataset/office/webcam/images/desktop_computer/frame_0008.jpg 8
414 | ../../dataset/office/webcam/images/desktop_computer/frame_0009.jpg 8
415 | ../../dataset/office/webcam/images/desktop_computer/frame_0010.jpg 8
416 | ../../dataset/office/webcam/images/desktop_computer/frame_0011.jpg 8
417 | ../../dataset/office/webcam/images/desktop_computer/frame_0012.jpg 8
418 | ../../dataset/office/webcam/images/desktop_computer/frame_0013.jpg 8
419 | ../../dataset/office/webcam/images/desktop_computer/frame_0014.jpg 8
420 | ../../dataset/office/webcam/images/desktop_computer/frame_0015.jpg 8
421 | ../../dataset/office/webcam/images/desktop_computer/frame_0016.jpg 8
422 | ../../dataset/office/webcam/images/desktop_computer/frame_0017.jpg 8
423 | ../../dataset/office/webcam/images/desktop_computer/frame_0018.jpg 8
424 | ../../dataset/office/webcam/images/desktop_computer/frame_0019.jpg 8
425 | ../../dataset/office/webcam/images/desktop_computer/frame_0020.jpg 8
426 | ../../dataset/office/webcam/images/desktop_computer/frame_0021.jpg 8
427 | ../../dataset/office/webcam/images/speaker/frame_0001.jpg 27
428 | ../../dataset/office/webcam/images/speaker/frame_0002.jpg 27
429 | ../../dataset/office/webcam/images/speaker/frame_0003.jpg 27
430 | ../../dataset/office/webcam/images/speaker/frame_0004.jpg 27
431 | ../../dataset/office/webcam/images/speaker/frame_0005.jpg 27
432 | ../../dataset/office/webcam/images/speaker/frame_0006.jpg 27
433 | ../../dataset/office/webcam/images/speaker/frame_0007.jpg 27
434 | ../../dataset/office/webcam/images/speaker/frame_0008.jpg 27
435 | ../../dataset/office/webcam/images/speaker/frame_0009.jpg 27
436 | ../../dataset/office/webcam/images/speaker/frame_0010.jpg 27
437 | ../../dataset/office/webcam/images/speaker/frame_0011.jpg 27
438 | ../../dataset/office/webcam/images/speaker/frame_0012.jpg 27
439 | ../../dataset/office/webcam/images/speaker/frame_0013.jpg 27
440 | ../../dataset/office/webcam/images/speaker/frame_0014.jpg 27
441 | ../../dataset/office/webcam/images/speaker/frame_0015.jpg 27
442 | ../../dataset/office/webcam/images/speaker/frame_0016.jpg 27
443 | ../../dataset/office/webcam/images/speaker/frame_0017.jpg 27
444 | ../../dataset/office/webcam/images/speaker/frame_0018.jpg 27
445 | ../../dataset/office/webcam/images/speaker/frame_0019.jpg 27
446 | ../../dataset/office/webcam/images/speaker/frame_0020.jpg 27
447 | ../../dataset/office/webcam/images/speaker/frame_0021.jpg 27
448 | ../../dataset/office/webcam/images/speaker/frame_0022.jpg 27
449 | ../../dataset/office/webcam/images/speaker/frame_0023.jpg 27
450 | ../../dataset/office/webcam/images/speaker/frame_0024.jpg 27
451 | ../../dataset/office/webcam/images/speaker/frame_0025.jpg 27
452 | ../../dataset/office/webcam/images/speaker/frame_0026.jpg 27
453 | ../../dataset/office/webcam/images/speaker/frame_0027.jpg 27
454 | ../../dataset/office/webcam/images/speaker/frame_0028.jpg 27
455 | ../../dataset/office/webcam/images/speaker/frame_0029.jpg 27
456 | ../../dataset/office/webcam/images/speaker/frame_0030.jpg 27
457 | ../../dataset/office/webcam/images/mobile_phone/frame_0001.jpg 14
458 | ../../dataset/office/webcam/images/mobile_phone/frame_0002.jpg 14
459 | ../../dataset/office/webcam/images/mobile_phone/frame_0003.jpg 14
460 | ../../dataset/office/webcam/images/mobile_phone/frame_0004.jpg 14
461 | ../../dataset/office/webcam/images/mobile_phone/frame_0005.jpg 14
462 | ../../dataset/office/webcam/images/mobile_phone/frame_0006.jpg 14
463 | ../../dataset/office/webcam/images/mobile_phone/frame_0007.jpg 14
464 | ../../dataset/office/webcam/images/mobile_phone/frame_0008.jpg 14
465 | ../../dataset/office/webcam/images/mobile_phone/frame_0009.jpg 14
466 | ../../dataset/office/webcam/images/mobile_phone/frame_0010.jpg 14
467 | ../../dataset/office/webcam/images/mobile_phone/frame_0011.jpg 14
468 | ../../dataset/office/webcam/images/mobile_phone/frame_0012.jpg 14
469 | ../../dataset/office/webcam/images/mobile_phone/frame_0013.jpg 14
470 | ../../dataset/office/webcam/images/mobile_phone/frame_0014.jpg 14
471 | ../../dataset/office/webcam/images/mobile_phone/frame_0015.jpg 14
472 | ../../dataset/office/webcam/images/mobile_phone/frame_0016.jpg 14
473 | ../../dataset/office/webcam/images/mobile_phone/frame_0017.jpg 14
474 | ../../dataset/office/webcam/images/mobile_phone/frame_0018.jpg 14
475 | ../../dataset/office/webcam/images/mobile_phone/frame_0019.jpg 14
476 | ../../dataset/office/webcam/images/mobile_phone/frame_0020.jpg 14
477 | ../../dataset/office/webcam/images/mobile_phone/frame_0021.jpg 14
478 | ../../dataset/office/webcam/images/mobile_phone/frame_0022.jpg 14
479 | ../../dataset/office/webcam/images/mobile_phone/frame_0023.jpg 14
480 | ../../dataset/office/webcam/images/mobile_phone/frame_0024.jpg 14
481 | ../../dataset/office/webcam/images/mobile_phone/frame_0025.jpg 14
482 | ../../dataset/office/webcam/images/mobile_phone/frame_0026.jpg 14
483 | ../../dataset/office/webcam/images/mobile_phone/frame_0027.jpg 14
484 | ../../dataset/office/webcam/images/mobile_phone/frame_0028.jpg 14
485 | ../../dataset/office/webcam/images/mobile_phone/frame_0029.jpg 14
486 | ../../dataset/office/webcam/images/mobile_phone/frame_0030.jpg 14
487 | ../../dataset/office/webcam/images/paper_notebook/frame_0001.jpg 18
488 | ../../dataset/office/webcam/images/paper_notebook/frame_0002.jpg 18
489 | ../../dataset/office/webcam/images/paper_notebook/frame_0003.jpg 18
490 | ../../dataset/office/webcam/images/paper_notebook/frame_0004.jpg 18
491 | ../../dataset/office/webcam/images/paper_notebook/frame_0005.jpg 18
492 | ../../dataset/office/webcam/images/paper_notebook/frame_0006.jpg 18
493 | ../../dataset/office/webcam/images/paper_notebook/frame_0007.jpg 18
494 | ../../dataset/office/webcam/images/paper_notebook/frame_0008.jpg 18
495 | ../../dataset/office/webcam/images/paper_notebook/frame_0009.jpg 18
496 | ../../dataset/office/webcam/images/paper_notebook/frame_0010.jpg 18
497 | ../../dataset/office/webcam/images/paper_notebook/frame_0011.jpg 18
498 | ../../dataset/office/webcam/images/paper_notebook/frame_0012.jpg 18
499 | ../../dataset/office/webcam/images/paper_notebook/frame_0013.jpg 18
500 | ../../dataset/office/webcam/images/paper_notebook/frame_0014.jpg 18
501 | ../../dataset/office/webcam/images/paper_notebook/frame_0015.jpg 18
502 | ../../dataset/office/webcam/images/paper_notebook/frame_0016.jpg 18
503 | ../../dataset/office/webcam/images/paper_notebook/frame_0017.jpg 18
504 | ../../dataset/office/webcam/images/paper_notebook/frame_0018.jpg 18
505 | ../../dataset/office/webcam/images/paper_notebook/frame_0019.jpg 18
506 | ../../dataset/office/webcam/images/paper_notebook/frame_0020.jpg 18
507 | ../../dataset/office/webcam/images/paper_notebook/frame_0021.jpg 18
508 | ../../dataset/office/webcam/images/paper_notebook/frame_0022.jpg 18
509 | ../../dataset/office/webcam/images/paper_notebook/frame_0023.jpg 18
510 | ../../dataset/office/webcam/images/paper_notebook/frame_0024.jpg 18
511 | ../../dataset/office/webcam/images/paper_notebook/frame_0025.jpg 18
512 | ../../dataset/office/webcam/images/paper_notebook/frame_0026.jpg 18
513 | ../../dataset/office/webcam/images/paper_notebook/frame_0027.jpg 18
514 | ../../dataset/office/webcam/images/paper_notebook/frame_0028.jpg 18
515 | ../../dataset/office/webcam/images/ruler/frame_0001.jpg 25
516 | ../../dataset/office/webcam/images/ruler/frame_0002.jpg 25
517 | ../../dataset/office/webcam/images/ruler/frame_0003.jpg 25
518 | ../../dataset/office/webcam/images/ruler/frame_0004.jpg 25
519 | ../../dataset/office/webcam/images/ruler/frame_0005.jpg 25
520 | ../../dataset/office/webcam/images/ruler/frame_0006.jpg 25
521 | ../../dataset/office/webcam/images/ruler/frame_0007.jpg 25
522 | ../../dataset/office/webcam/images/ruler/frame_0008.jpg 25
523 | ../../dataset/office/webcam/images/ruler/frame_0009.jpg 25
524 | ../../dataset/office/webcam/images/ruler/frame_0010.jpg 25
525 | ../../dataset/office/webcam/images/ruler/frame_0011.jpg 25
526 | ../../dataset/office/webcam/images/letter_tray/frame_0001.jpg 13
527 | ../../dataset/office/webcam/images/letter_tray/frame_0002.jpg 13
528 | ../../dataset/office/webcam/images/letter_tray/frame_0003.jpg 13
529 | ../../dataset/office/webcam/images/letter_tray/frame_0004.jpg 13
530 | ../../dataset/office/webcam/images/letter_tray/frame_0005.jpg 13
531 | ../../dataset/office/webcam/images/letter_tray/frame_0006.jpg 13
532 | ../../dataset/office/webcam/images/letter_tray/frame_0007.jpg 13
533 | ../../dataset/office/webcam/images/letter_tray/frame_0008.jpg 13
534 | ../../dataset/office/webcam/images/letter_tray/frame_0009.jpg 13
535 | ../../dataset/office/webcam/images/letter_tray/frame_0010.jpg 13
536 | ../../dataset/office/webcam/images/letter_tray/frame_0011.jpg 13
537 | ../../dataset/office/webcam/images/letter_tray/frame_0012.jpg 13
538 | ../../dataset/office/webcam/images/letter_tray/frame_0013.jpg 13
539 | ../../dataset/office/webcam/images/letter_tray/frame_0014.jpg 13
540 | ../../dataset/office/webcam/images/letter_tray/frame_0015.jpg 13
541 | ../../dataset/office/webcam/images/letter_tray/frame_0016.jpg 13
542 | ../../dataset/office/webcam/images/letter_tray/frame_0017.jpg 13
543 | ../../dataset/office/webcam/images/letter_tray/frame_0018.jpg 13
544 | ../../dataset/office/webcam/images/letter_tray/frame_0019.jpg 13
545 | ../../dataset/office/webcam/images/file_cabinet/frame_0001.jpg 9
546 | ../../dataset/office/webcam/images/file_cabinet/frame_0002.jpg 9
547 | ../../dataset/office/webcam/images/file_cabinet/frame_0003.jpg 9
548 | ../../dataset/office/webcam/images/file_cabinet/frame_0004.jpg 9
549 | ../../dataset/office/webcam/images/file_cabinet/frame_0005.jpg 9
550 | ../../dataset/office/webcam/images/file_cabinet/frame_0006.jpg 9
551 | ../../dataset/office/webcam/images/file_cabinet/frame_0007.jpg 9
552 | ../../dataset/office/webcam/images/file_cabinet/frame_0008.jpg 9
553 | ../../dataset/office/webcam/images/file_cabinet/frame_0009.jpg 9
554 | ../../dataset/office/webcam/images/file_cabinet/frame_0010.jpg 9
555 | ../../dataset/office/webcam/images/file_cabinet/frame_0011.jpg 9
556 | ../../dataset/office/webcam/images/file_cabinet/frame_0012.jpg 9
557 | ../../dataset/office/webcam/images/file_cabinet/frame_0013.jpg 9
558 | ../../dataset/office/webcam/images/file_cabinet/frame_0014.jpg 9
559 | ../../dataset/office/webcam/images/file_cabinet/frame_0015.jpg 9
560 | ../../dataset/office/webcam/images/file_cabinet/frame_0016.jpg 9
561 | ../../dataset/office/webcam/images/file_cabinet/frame_0017.jpg 9
562 | ../../dataset/office/webcam/images/file_cabinet/frame_0018.jpg 9
563 | ../../dataset/office/webcam/images/file_cabinet/frame_0019.jpg 9
564 | ../../dataset/office/webcam/images/phone/frame_0001.jpg 20
565 | ../../dataset/office/webcam/images/phone/frame_0002.jpg 20
566 | ../../dataset/office/webcam/images/phone/frame_0003.jpg 20
567 | ../../dataset/office/webcam/images/phone/frame_0004.jpg 20
568 | ../../dataset/office/webcam/images/phone/frame_0005.jpg 20
569 | ../../dataset/office/webcam/images/phone/frame_0006.jpg 20
570 | ../../dataset/office/webcam/images/phone/frame_0007.jpg 20
571 | ../../dataset/office/webcam/images/phone/frame_0008.jpg 20
572 | ../../dataset/office/webcam/images/phone/frame_0009.jpg 20
573 | ../../dataset/office/webcam/images/phone/frame_0010.jpg 20
574 | ../../dataset/office/webcam/images/phone/frame_0011.jpg 20
575 | ../../dataset/office/webcam/images/phone/frame_0012.jpg 20
576 | ../../dataset/office/webcam/images/phone/frame_0013.jpg 20
577 | ../../dataset/office/webcam/images/phone/frame_0014.jpg 20
578 | ../../dataset/office/webcam/images/phone/frame_0015.jpg 20
579 | ../../dataset/office/webcam/images/phone/frame_0016.jpg 20
580 | ../../dataset/office/webcam/images/bookcase/frame_0001.jpg 3
581 | ../../dataset/office/webcam/images/bookcase/frame_0002.jpg 3
582 | ../../dataset/office/webcam/images/bookcase/frame_0003.jpg 3
583 | ../../dataset/office/webcam/images/bookcase/frame_0004.jpg 3
584 | ../../dataset/office/webcam/images/bookcase/frame_0005.jpg 3
585 | ../../dataset/office/webcam/images/bookcase/frame_0006.jpg 3
586 | ../../dataset/office/webcam/images/bookcase/frame_0007.jpg 3
587 | ../../dataset/office/webcam/images/bookcase/frame_0008.jpg 3
588 | ../../dataset/office/webcam/images/bookcase/frame_0009.jpg 3
589 | ../../dataset/office/webcam/images/bookcase/frame_0010.jpg 3
590 | ../../dataset/office/webcam/images/bookcase/frame_0011.jpg 3
591 | ../../dataset/office/webcam/images/bookcase/frame_0012.jpg 3
592 | ../../dataset/office/webcam/images/projector/frame_0001.jpg 22
593 | ../../dataset/office/webcam/images/projector/frame_0002.jpg 22
594 | ../../dataset/office/webcam/images/projector/frame_0003.jpg 22
595 | ../../dataset/office/webcam/images/projector/frame_0004.jpg 22
596 | ../../dataset/office/webcam/images/projector/frame_0005.jpg 22
597 | ../../dataset/office/webcam/images/projector/frame_0006.jpg 22
598 | ../../dataset/office/webcam/images/projector/frame_0007.jpg 22
599 | ../../dataset/office/webcam/images/projector/frame_0008.jpg 22
600 | ../../dataset/office/webcam/images/projector/frame_0009.jpg 22
601 | ../../dataset/office/webcam/images/projector/frame_0010.jpg 22
602 | ../../dataset/office/webcam/images/projector/frame_0011.jpg 22
603 | ../../dataset/office/webcam/images/projector/frame_0012.jpg 22
604 | ../../dataset/office/webcam/images/projector/frame_0013.jpg 22
605 | ../../dataset/office/webcam/images/projector/frame_0014.jpg 22
606 | ../../dataset/office/webcam/images/projector/frame_0015.jpg 22
607 | ../../dataset/office/webcam/images/projector/frame_0016.jpg 22
608 | ../../dataset/office/webcam/images/projector/frame_0017.jpg 22
609 | ../../dataset/office/webcam/images/projector/frame_0018.jpg 22
610 | ../../dataset/office/webcam/images/projector/frame_0019.jpg 22
611 | ../../dataset/office/webcam/images/projector/frame_0020.jpg 22
612 | ../../dataset/office/webcam/images/projector/frame_0021.jpg 22
613 | ../../dataset/office/webcam/images/projector/frame_0022.jpg 22
614 | ../../dataset/office/webcam/images/projector/frame_0023.jpg 22
615 | ../../dataset/office/webcam/images/projector/frame_0024.jpg 22
616 | ../../dataset/office/webcam/images/projector/frame_0025.jpg 22
617 | ../../dataset/office/webcam/images/projector/frame_0026.jpg 22
618 | ../../dataset/office/webcam/images/projector/frame_0027.jpg 22
619 | ../../dataset/office/webcam/images/projector/frame_0028.jpg 22
620 | ../../dataset/office/webcam/images/projector/frame_0029.jpg 22
621 | ../../dataset/office/webcam/images/projector/frame_0030.jpg 22
622 | ../../dataset/office/webcam/images/stapler/frame_0001.jpg 28
623 | ../../dataset/office/webcam/images/stapler/frame_0002.jpg 28
624 | ../../dataset/office/webcam/images/stapler/frame_0003.jpg 28
625 | ../../dataset/office/webcam/images/stapler/frame_0004.jpg 28
626 | ../../dataset/office/webcam/images/stapler/frame_0005.jpg 28
627 | ../../dataset/office/webcam/images/stapler/frame_0006.jpg 28
628 | ../../dataset/office/webcam/images/stapler/frame_0007.jpg 28
629 | ../../dataset/office/webcam/images/stapler/frame_0008.jpg 28
630 | ../../dataset/office/webcam/images/stapler/frame_0009.jpg 28
631 | ../../dataset/office/webcam/images/stapler/frame_0010.jpg 28
632 | ../../dataset/office/webcam/images/stapler/frame_0011.jpg 28
633 | ../../dataset/office/webcam/images/stapler/frame_0012.jpg 28
634 | ../../dataset/office/webcam/images/stapler/frame_0013.jpg 28
635 | ../../dataset/office/webcam/images/stapler/frame_0014.jpg 28
636 | ../../dataset/office/webcam/images/stapler/frame_0015.jpg 28
637 | ../../dataset/office/webcam/images/stapler/frame_0016.jpg 28
638 | ../../dataset/office/webcam/images/stapler/frame_0017.jpg 28
639 | ../../dataset/office/webcam/images/stapler/frame_0018.jpg 28
640 | ../../dataset/office/webcam/images/stapler/frame_0019.jpg 28
641 | ../../dataset/office/webcam/images/stapler/frame_0020.jpg 28
642 | ../../dataset/office/webcam/images/stapler/frame_0021.jpg 28
643 | ../../dataset/office/webcam/images/stapler/frame_0022.jpg 28
644 | ../../dataset/office/webcam/images/stapler/frame_0023.jpg 28
645 | ../../dataset/office/webcam/images/stapler/frame_0024.jpg 28
646 | ../../dataset/office/webcam/images/trash_can/frame_0001.jpg 30
647 | ../../dataset/office/webcam/images/trash_can/frame_0002.jpg 30
648 | ../../dataset/office/webcam/images/trash_can/frame_0003.jpg 30
649 | ../../dataset/office/webcam/images/trash_can/frame_0004.jpg 30
650 | ../../dataset/office/webcam/images/trash_can/frame_0005.jpg 30
651 | ../../dataset/office/webcam/images/trash_can/frame_0006.jpg 30
652 | ../../dataset/office/webcam/images/trash_can/frame_0007.jpg 30
653 | ../../dataset/office/webcam/images/trash_can/frame_0008.jpg 30
654 | ../../dataset/office/webcam/images/trash_can/frame_0009.jpg 30
655 | ../../dataset/office/webcam/images/trash_can/frame_0010.jpg 30
656 | ../../dataset/office/webcam/images/trash_can/frame_0011.jpg 30
657 | ../../dataset/office/webcam/images/trash_can/frame_0012.jpg 30
658 | ../../dataset/office/webcam/images/trash_can/frame_0013.jpg 30
659 | ../../dataset/office/webcam/images/trash_can/frame_0014.jpg 30
660 | ../../dataset/office/webcam/images/trash_can/frame_0015.jpg 30
661 | ../../dataset/office/webcam/images/trash_can/frame_0016.jpg 30
662 | ../../dataset/office/webcam/images/trash_can/frame_0017.jpg 30
663 | ../../dataset/office/webcam/images/trash_can/frame_0018.jpg 30
664 | ../../dataset/office/webcam/images/trash_can/frame_0019.jpg 30
665 | ../../dataset/office/webcam/images/trash_can/frame_0020.jpg 30
666 | ../../dataset/office/webcam/images/trash_can/frame_0021.jpg 30
667 | ../../dataset/office/webcam/images/bike_helmet/frame_0001.jpg 2
668 | ../../dataset/office/webcam/images/bike_helmet/frame_0002.jpg 2
669 | ../../dataset/office/webcam/images/bike_helmet/frame_0003.jpg 2
670 | ../../dataset/office/webcam/images/bike_helmet/frame_0004.jpg 2
671 | ../../dataset/office/webcam/images/bike_helmet/frame_0005.jpg 2
672 | ../../dataset/office/webcam/images/bike_helmet/frame_0006.jpg 2
673 | ../../dataset/office/webcam/images/bike_helmet/frame_0007.jpg 2
674 | ../../dataset/office/webcam/images/bike_helmet/frame_0008.jpg 2
675 | ../../dataset/office/webcam/images/bike_helmet/frame_0009.jpg 2
676 | ../../dataset/office/webcam/images/bike_helmet/frame_0010.jpg 2
677 | ../../dataset/office/webcam/images/bike_helmet/frame_0011.jpg 2
678 | ../../dataset/office/webcam/images/bike_helmet/frame_0012.jpg 2
679 | ../../dataset/office/webcam/images/bike_helmet/frame_0013.jpg 2
680 | ../../dataset/office/webcam/images/bike_helmet/frame_0014.jpg 2
681 | ../../dataset/office/webcam/images/bike_helmet/frame_0015.jpg 2
682 | ../../dataset/office/webcam/images/bike_helmet/frame_0016.jpg 2
683 | ../../dataset/office/webcam/images/bike_helmet/frame_0017.jpg 2
684 | ../../dataset/office/webcam/images/bike_helmet/frame_0018.jpg 2
685 | ../../dataset/office/webcam/images/bike_helmet/frame_0019.jpg 2
686 | ../../dataset/office/webcam/images/bike_helmet/frame_0020.jpg 2
687 | ../../dataset/office/webcam/images/bike_helmet/frame_0021.jpg 2
688 | ../../dataset/office/webcam/images/bike_helmet/frame_0022.jpg 2
689 | ../../dataset/office/webcam/images/bike_helmet/frame_0023.jpg 2
690 | ../../dataset/office/webcam/images/bike_helmet/frame_0024.jpg 2
691 | ../../dataset/office/webcam/images/bike_helmet/frame_0025.jpg 2
692 | ../../dataset/office/webcam/images/bike_helmet/frame_0026.jpg 2
693 | ../../dataset/office/webcam/images/bike_helmet/frame_0027.jpg 2
694 | ../../dataset/office/webcam/images/bike_helmet/frame_0028.jpg 2
695 | ../../dataset/office/webcam/images/headphones/frame_0001.jpg 10
696 | ../../dataset/office/webcam/images/headphones/frame_0002.jpg 10
697 | ../../dataset/office/webcam/images/headphones/frame_0003.jpg 10
698 | ../../dataset/office/webcam/images/headphones/frame_0004.jpg 10
699 | ../../dataset/office/webcam/images/headphones/frame_0005.jpg 10
700 | ../../dataset/office/webcam/images/headphones/frame_0006.jpg 10
701 | ../../dataset/office/webcam/images/headphones/frame_0007.jpg 10
702 | ../../dataset/office/webcam/images/headphones/frame_0008.jpg 10
703 | ../../dataset/office/webcam/images/headphones/frame_0009.jpg 10
704 | ../../dataset/office/webcam/images/headphones/frame_0010.jpg 10
705 | ../../dataset/office/webcam/images/headphones/frame_0011.jpg 10
706 | ../../dataset/office/webcam/images/headphones/frame_0012.jpg 10
707 | ../../dataset/office/webcam/images/headphones/frame_0013.jpg 10
708 | ../../dataset/office/webcam/images/headphones/frame_0014.jpg 10
709 | ../../dataset/office/webcam/images/headphones/frame_0015.jpg 10
710 | ../../dataset/office/webcam/images/headphones/frame_0016.jpg 10
711 | ../../dataset/office/webcam/images/headphones/frame_0017.jpg 10
712 | ../../dataset/office/webcam/images/headphones/frame_0018.jpg 10
713 | ../../dataset/office/webcam/images/headphones/frame_0019.jpg 10
714 | ../../dataset/office/webcam/images/headphones/frame_0020.jpg 10
715 | ../../dataset/office/webcam/images/headphones/frame_0021.jpg 10
716 | ../../dataset/office/webcam/images/headphones/frame_0022.jpg 10
717 | ../../dataset/office/webcam/images/headphones/frame_0023.jpg 10
718 | ../../dataset/office/webcam/images/headphones/frame_0024.jpg 10
719 | ../../dataset/office/webcam/images/headphones/frame_0025.jpg 10
720 | ../../dataset/office/webcam/images/headphones/frame_0026.jpg 10
721 | ../../dataset/office/webcam/images/headphones/frame_0027.jpg 10
722 | ../../dataset/office/webcam/images/desk_lamp/frame_0001.jpg 7
723 | ../../dataset/office/webcam/images/desk_lamp/frame_0002.jpg 7
724 | ../../dataset/office/webcam/images/desk_lamp/frame_0003.jpg 7
725 | ../../dataset/office/webcam/images/desk_lamp/frame_0004.jpg 7
726 | ../../dataset/office/webcam/images/desk_lamp/frame_0005.jpg 7
727 | ../../dataset/office/webcam/images/desk_lamp/frame_0006.jpg 7
728 | ../../dataset/office/webcam/images/desk_lamp/frame_0007.jpg 7
729 | ../../dataset/office/webcam/images/desk_lamp/frame_0008.jpg 7
730 | ../../dataset/office/webcam/images/desk_lamp/frame_0009.jpg 7
731 | ../../dataset/office/webcam/images/desk_lamp/frame_0010.jpg 7
732 | ../../dataset/office/webcam/images/desk_lamp/frame_0011.jpg 7
733 | ../../dataset/office/webcam/images/desk_lamp/frame_0012.jpg 7
734 | ../../dataset/office/webcam/images/desk_lamp/frame_0013.jpg 7
735 | ../../dataset/office/webcam/images/desk_lamp/frame_0014.jpg 7
736 | ../../dataset/office/webcam/images/desk_lamp/frame_0015.jpg 7
737 | ../../dataset/office/webcam/images/desk_lamp/frame_0016.jpg 7
738 | ../../dataset/office/webcam/images/desk_lamp/frame_0017.jpg 7
739 | ../../dataset/office/webcam/images/desk_lamp/frame_0018.jpg 7
740 | ../../dataset/office/webcam/images/desk_chair/frame_0001.jpg 6
741 | ../../dataset/office/webcam/images/desk_chair/frame_0002.jpg 6
742 | ../../dataset/office/webcam/images/desk_chair/frame_0003.jpg 6
743 | ../../dataset/office/webcam/images/desk_chair/frame_0004.jpg 6
744 | ../../dataset/office/webcam/images/desk_chair/frame_0005.jpg 6
745 | ../../dataset/office/webcam/images/desk_chair/frame_0006.jpg 6
746 | ../../dataset/office/webcam/images/desk_chair/frame_0007.jpg 6
747 | ../../dataset/office/webcam/images/desk_chair/frame_0008.jpg 6
748 | ../../dataset/office/webcam/images/desk_chair/frame_0009.jpg 6
749 | ../../dataset/office/webcam/images/desk_chair/frame_0010.jpg 6
750 | ../../dataset/office/webcam/images/desk_chair/frame_0011.jpg 6
751 | ../../dataset/office/webcam/images/desk_chair/frame_0012.jpg 6
752 | ../../dataset/office/webcam/images/desk_chair/frame_0013.jpg 6
753 | ../../dataset/office/webcam/images/desk_chair/frame_0014.jpg 6
754 | ../../dataset/office/webcam/images/desk_chair/frame_0015.jpg 6
755 | ../../dataset/office/webcam/images/desk_chair/frame_0016.jpg 6
756 | ../../dataset/office/webcam/images/desk_chair/frame_0017.jpg 6
757 | ../../dataset/office/webcam/images/desk_chair/frame_0018.jpg 6
758 | ../../dataset/office/webcam/images/desk_chair/frame_0019.jpg 6
759 | ../../dataset/office/webcam/images/desk_chair/frame_0020.jpg 6
760 | ../../dataset/office/webcam/images/desk_chair/frame_0021.jpg 6
761 | ../../dataset/office/webcam/images/desk_chair/frame_0022.jpg 6
762 | ../../dataset/office/webcam/images/desk_chair/frame_0023.jpg 6
763 | ../../dataset/office/webcam/images/desk_chair/frame_0024.jpg 6
764 | ../../dataset/office/webcam/images/desk_chair/frame_0025.jpg 6
765 | ../../dataset/office/webcam/images/desk_chair/frame_0026.jpg 6
766 | ../../dataset/office/webcam/images/desk_chair/frame_0027.jpg 6
767 | ../../dataset/office/webcam/images/desk_chair/frame_0028.jpg 6
768 | ../../dataset/office/webcam/images/desk_chair/frame_0029.jpg 6
769 | ../../dataset/office/webcam/images/desk_chair/frame_0030.jpg 6
770 | ../../dataset/office/webcam/images/desk_chair/frame_0031.jpg 6
771 | ../../dataset/office/webcam/images/desk_chair/frame_0032.jpg 6
772 | ../../dataset/office/webcam/images/desk_chair/frame_0033.jpg 6
773 | ../../dataset/office/webcam/images/desk_chair/frame_0034.jpg 6
774 | ../../dataset/office/webcam/images/desk_chair/frame_0035.jpg 6
775 | ../../dataset/office/webcam/images/desk_chair/frame_0036.jpg 6
776 | ../../dataset/office/webcam/images/desk_chair/frame_0037.jpg 6
777 | ../../dataset/office/webcam/images/desk_chair/frame_0038.jpg 6
778 | ../../dataset/office/webcam/images/desk_chair/frame_0039.jpg 6
779 | ../../dataset/office/webcam/images/desk_chair/frame_0040.jpg 6
780 | ../../dataset/office/webcam/images/bottle/frame_0001.jpg 4
781 | ../../dataset/office/webcam/images/bottle/frame_0002.jpg 4
782 | ../../dataset/office/webcam/images/bottle/frame_0003.jpg 4
783 | ../../dataset/office/webcam/images/bottle/frame_0004.jpg 4
784 | ../../dataset/office/webcam/images/bottle/frame_0005.jpg 4
785 | ../../dataset/office/webcam/images/bottle/frame_0006.jpg 4
786 | ../../dataset/office/webcam/images/bottle/frame_0007.jpg 4
787 | ../../dataset/office/webcam/images/bottle/frame_0008.jpg 4
788 | ../../dataset/office/webcam/images/bottle/frame_0009.jpg 4
789 | ../../dataset/office/webcam/images/bottle/frame_0010.jpg 4
790 | ../../dataset/office/webcam/images/bottle/frame_0011.jpg 4
791 | ../../dataset/office/webcam/images/bottle/frame_0012.jpg 4
792 | ../../dataset/office/webcam/images/bottle/frame_0013.jpg 4
793 | ../../dataset/office/webcam/images/bottle/frame_0014.jpg 4
794 | ../../dataset/office/webcam/images/bottle/frame_0015.jpg 4
795 | ../../dataset/office/webcam/images/bottle/frame_0016.jpg 4
796 |
--------------------------------------------------------------------------------
/dataset.py:
--------------------------------------------------------------------------------
1 | import cv2
2 | import numpy as np
3 | from torch.utils import data
4 | import torchvision.transforms as tv
5 |
6 |
7 | class Office(data.Dataset):
8 |
9 | def __init__(self, list, training=True):
10 | self.images = []
11 | self.labels = []
12 | self.multi_scale = [256, 257]
13 | self.output_size = [227, 227]
14 | self.training = training
15 | self.mean_color=[104.006,116.668,122.678]
16 |
17 | list_file = open(list)
18 | lines = list_file.readlines()
19 | for line in lines:
20 | fields = line.split()
21 | self.images.append(fields[0])
22 | self.labels.append(int(fields[1]))
23 |
24 | def __len__(self):
25 | return len(self.images)
26 |
27 | def __getitem__(self, index):
28 | image_path = self.images[index]
29 | label = self.labels[index]
30 | img = cv2.imread(image_path)
31 | if type(img) == None:
32 | print('Error: Image at {} not found.'.format(image_path))
33 |
34 | if self.training and np.random.random() < 0.5:
35 | img = cv2.flip(img, 1)
36 | new_size = np.random.randint(self.multi_scale[0], self.multi_scale[1], 1)[0]
37 |
38 | img = cv2.resize(img, (new_size, new_size))
39 | img = img.astype(np.float32)
40 |
41 | # cropping
42 | if self.training:
43 | diff = new_size - self.output_size[0]
44 | offset_x = np.random.randint(0, diff, 1)[0]
45 | offset_y = np.random.randint(0, diff, 1)[0]
46 | else:
47 | offset_x = img.shape[0]//2 - self.output_size[0] // 2
48 | offset_y = img.shape[1]//2 - self.output_size[1] // 2
49 |
50 | img = img[offset_x:(offset_x+self.output_size[0]),
51 | offset_y:(offset_y+self.output_size[1])]
52 |
53 | # substract mean
54 | img -= np.array(self.mean_color)
55 | # img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
56 |
57 | # ToTensor transform cv2 HWC->CHW, only byteTensor will be div by 255.
58 | tensor = tv.ToTensor()
59 | img = tensor(img)
60 | # img = np.transpose(img, (2, 0, 1))
61 |
62 | return img, label
--------------------------------------------------------------------------------
/model.py:
--------------------------------------------------------------------------------
1 | import itertools
2 | from itertools import chain
3 | import torch
4 | import torch.nn as nn
5 | import utils
6 | from utils import LRN
7 |
8 | class AlexNet(nn.Module):
9 |
10 | def __init__(self, cudable, n_class):
11 | super(AlexNet, self).__init__()
12 | self.cudable = cudable
13 | self.n_class = n_class
14 | self.decay = 0.3
15 | self.s_centroid = torch.zeros(self.n_class, 256)
16 | self.t_centroid = torch.zeros(self.n_class, 256)
17 | if self.cudable:
18 | self.s_centroid = self.s_centroid.cuda()
19 | self.t_centroid = self.t_centroid.cuda()
20 | self.features = nn.Sequential(
21 | nn.Conv2d(3, 96, 11, stride=4),
22 | nn.ReLU(inplace=True),
23 | nn.MaxPool2d(3, stride=2),
24 | # nn.LocalResponseNorm(3, alpha=1e-5),
25 | LRN(local_size=5, alpha=1e-4, beta=0.75),
26 | nn.Conv2d(96, 256, 5, stride=1, padding=2, groups=2),
27 | nn.ReLU(inplace=True),
28 | nn.MaxPool2d(3, stride=2),
29 | LRN(local_size=5, alpha=1e-4, beta=0.75),
30 | # nn.LocalResponseNorm(3, alpha=1e-5),
31 | nn.Conv2d(256, 384, 3, stride=1, padding=1),
32 | nn.ReLU(inplace=True),
33 | nn.Conv2d(384, 384, 3, stride=1, padding=1, groups=2),
34 | nn.ReLU(inplace=True),
35 | nn.Conv2d(384, 256, 3, stride=1, padding=1, groups=2),
36 | nn.ReLU(inplace=True),
37 | nn.MaxPool2d(3, stride=2)
38 | )
39 | self.classifier = nn.Sequential(
40 | nn.Linear(6*6*256, 4096),
41 | nn.ReLU(inplace=True),
42 | nn.Dropout(),
43 | nn.Linear(4096, 4096),
44 | nn.ReLU(inplace=True),
45 | nn.Dropout(),
46 | )
47 | self.fc8 = nn.Sequential(
48 | nn.Linear(4096, 256)
49 | )
50 | self.fc9 = nn.Sequential(
51 | nn.Linear(256, self.n_class)
52 | )
53 | self.softmax = nn.Softmax(dim=0)
54 | self.D = nn.Sequential(
55 | nn.Linear(256, 1024),
56 | nn.ReLU(inplace=True),
57 | nn.Dropout(),
58 | nn.Linear(1024, 1024),
59 | nn.ReLU(inplace=True),
60 | nn.Dropout(),
61 | nn.Linear(1024, 1)
62 | )
63 | self.init()
64 |
65 | def init(self):
66 | self.init_linear(self.fc8[0])
67 | self.init_linear(self.fc9[0], std=0.005)
68 | self.init_linear(self.D[0],D=True)
69 | self.init_linear(self.D[3],D=True)
70 | self.init_linear(self.D[6],D=True, std=0.3)
71 | self.CEloss, self.MSEloss, self.BCEloss = nn.CrossEntropyLoss(), nn.MSELoss(), nn.BCEWithLogitsLoss(reduction='mean')
72 | if self.cudable:
73 | self.CEloss, self.MSEloss, self.BCEloss = self.CEloss.cuda(), self.MSEloss.cuda(), self.BCEloss.cuda()
74 |
75 | def init_linear(self, m, std=0.01, D=False):
76 | # nn.init.normal_(m.weight.data, 0, std)
77 | # nn.init.xavier_normal_(m.weight)
78 | utils.truncated_normal_(m.weight.data, 0, std)
79 | if D:
80 | m.bias.data.fill_(0)
81 | else:
82 | m.bias.data.fill_(0.1)
83 |
84 | def forward(self, x, training=True):
85 | conv_out = self.features(x)
86 | flattened = conv_out.view(conv_out.size(0), -1)
87 | dense_out = self.classifier(flattened)
88 | feature = self.fc8(dense_out)
89 | score = self.fc9(feature)
90 | pred = self.softmax(score)
91 | return feature, score, pred
92 |
93 | def forward_D(self, feature):
94 | logit = self.D(feature)
95 | return logit
96 |
97 | def closs(self, y_pred, y):
98 | C_loss = self.CEloss(y_pred, y)
99 | return C_loss
100 |
101 | def adloss(self, s_logits, t_logits, s_feature, t_feature, y_s, y_t):
102 | n, d = s_feature.shape
103 |
104 | # get labels
105 | s_labels, t_labels = y_s, torch.max(y_t, 1)[1]
106 |
107 | # image number in each class
108 | ones = torch.ones_like(s_labels, dtype=torch.float)
109 | zeros = torch.zeros(self.n_class)
110 | if self.cudable:
111 | zeros = zeros.cuda()
112 | s_n_classes = zeros.scatter_add(0, s_labels, ones)
113 | t_n_classes = zeros.scatter_add(0, t_labels, ones)
114 |
115 | # image number cannot be 0, when calculating centroids
116 | ones = torch.ones_like(s_n_classes)
117 | s_n_classes = torch.max(s_n_classes, ones)
118 | t_n_classes = torch.max(t_n_classes, ones)
119 |
120 | # calculating centroids, sum and divide
121 | zeros = torch.zeros(self.n_class, d)
122 | if self.cudable:
123 | zeros = zeros.cuda()
124 | s_sum_feature = zeros.scatter_add(0, torch.transpose(s_labels.repeat(d, 1), 1, 0), s_feature)
125 | t_sum_feature = zeros.scatter_add(0, torch.transpose(t_labels.repeat(d, 1), 1, 0), t_feature)
126 | current_s_centroid = torch.div(s_sum_feature, s_n_classes.view(self.n_class, 1))
127 | current_t_centroid = torch.div(t_sum_feature, t_n_classes.view(self.n_class, 1))
128 |
129 | # Moving Centroid
130 | decay = self.decay
131 | s_centroid = (1-decay) * self.s_centroid + decay * current_s_centroid
132 | t_centroid = (1-decay) * self.t_centroid + decay * current_t_centroid
133 | semantic_loss = self.MSEloss(s_centroid, t_centroid)
134 | self.s_centroid = s_centroid.detach()
135 | self.t_centroid = t_centroid.detach()
136 |
137 | # sigmoid binary cross entropy with reduce mean
138 | D_real_loss = self.BCEloss(t_logits, torch.ones_like(t_logits))
139 | D_fake_loss = self.BCEloss(s_logits, torch.zeros_like(s_logits))
140 | D_loss = (D_real_loss + D_fake_loss) * 0.1
141 | G_loss = -D_loss
142 |
143 | return G_loss, D_loss, semantic_loss
144 |
145 | # To some extent, can be replaced by weight_decay param in optimizer.
146 | def regloss(self):
147 | Dregloss = [torch.sum(layer.weight ** 2) / 2 for layer in self.D if type(layer) == nn.Linear]
148 | layers = chain(self.features, self.classifier, self.fc8, self.fc9)
149 | Gregloss = [torch.sum(layer.weight ** 2) / 2 for layer in layers if type(layer) == nn.Conv2d or type(layer) == nn.Linear]
150 | mean = lambda x:0.0005 * torch.mean(torch.stack(x))
151 | return mean(Dregloss), mean(Gregloss)
152 |
153 |
154 | def get_optimizer(self, init_lr, lr_mult, lr_mult_D):
155 | w_finetune, b_finetune, w_train, b_train, w_D, b_D = [], [], [], [], [], []
156 |
157 | finetune_layers = itertools.chain(self.features, self.classifier)
158 | train_layers = itertools.chain(self.fc8, self.fc9)
159 | for layer in finetune_layers:
160 | if type(layer) == nn.Conv2d or type(layer) == nn.Linear:
161 | w_finetune.append(layer.weight)
162 | b_finetune.append(layer.bias)
163 | for layer in train_layers:
164 | if type(layer) == nn.Linear:
165 | w_train.append(layer.weight)
166 | b_train.append(layer.bias)
167 | for layer in self.D:
168 | if type(layer) == nn.Linear:
169 | w_D.append(layer.weight)
170 | b_D.append(layer.bias)
171 |
172 | opt = torch.optim.SGD([{'params': w_finetune, 'lr': init_lr * lr_mult[0]},
173 | {'params': b_finetune, 'lr': init_lr * lr_mult[1]},
174 | {'params': w_train, 'lr': init_lr * lr_mult[2]},
175 | {'params': b_train, 'lr': init_lr * lr_mult[3]}],
176 | lr=init_lr,momentum=0.9)
177 |
178 | opt_D = torch.optim.SGD([{'params': w_D, 'lr': init_lr * lr_mult_D[0]},
179 | {'params': b_D, 'lr': init_lr * lr_mult_D[1]}],
180 | lr=init_lr,momentum=0.9)
181 |
182 | return opt, opt_D
183 |
--------------------------------------------------------------------------------
/train.py:
--------------------------------------------------------------------------------
1 | import os
2 | import math
3 | import argparse
4 | import numpy as np
5 | import torch
6 | import torch.nn as nn
7 |
8 | import dataset
9 | from model import AlexNet
10 | import utils
11 |
12 |
13 | def lr_schedule(opt, epoch, mult):
14 | lr = init_lr / pow(1 + 0.001 * epoch, 0.75)
15 | for ind, param_group in enumerate(opt.param_groups):
16 | param_group['lr'] = lr * mult[ind]
17 | return lr
18 |
19 | def adaptation_factor(x):
20 | if x>= 1.0:
21 | return 1.0
22 | den = 1.0 + math.exp(-10 * x)
23 | lamb = 2.0 / den - 1.0
24 | return lamb
25 |
26 | def output(mes):
27 | print(mes)
28 | # log.write(mes)
29 |
30 |
31 | parser = argparse.ArgumentParser()
32 | parser.add_argument('--s', default=0, type=int)
33 | parser.add_argument('--t', default=1, type=int)
34 | parser.add_argument('--gpu', default=1, type=int)
35 | parser.add_argument('--resume', default='', type=str) # resume from pretrained bvlc_alexnet model
36 | parser.add_argument('--da', default=1, type=int) # 1 for doing domain adaptation
37 | args = parser.parse_args()
38 | resume = args.resume
39 | da = args.da
40 | os.environ['CUDA_VISIBLE_DEVICES'] = str(args.gpu)
41 | cuda = torch.cuda.is_available()
42 |
43 | init_lr = 1e-2
44 | batch_size = 100
45 | max_epoch = 10000
46 | lr_mult = [0.1, 0.2, 1, 2]
47 | lr_mult_D = [1, 2]
48 | dataset_names = ['amazon', 'webcam', 'dslr']
49 | s_name = dataset_names[args.s]
50 | t_name = dataset_names[args.t]
51 | s_list_path = './data_list/' + s_name + '_list.txt'
52 | t_list_path = './data_list/' + t_name + '_list.txt'
53 | s_folder_path = '../../dataset/office/' + s_name + '/images'
54 | t_folder_path = '../../dataset/office/' + t_name + '/images'
55 | n_class = 31
56 | log_name = '_'.join(str(a) for a in [s_name, t_name, str(batch_size), str(da), str(init_lr)])
57 | log = open('log/' + log_name + '.log', 'w')
58 | pretrain_path = 'checkpoint/' + resume + '.pth'
59 | checkpoint_save_path = 'checkpoint/' + log_name + '.pth'
60 | print('GPU: {}'.format(args.gpu))
61 | print('source: {}, target: {}, batch_size: {}, init_lr: {}'.format(s_name, t_name, batch_size, init_lr))
62 | print('lr_mult: {}, lr_mult_D: {}, resume: {}, da: {}'.format(lr_mult, lr_mult_D, resume, da))
63 |
64 | s_loader = torch.utils.data.DataLoader(dataset.Office(s_list_path),
65 | batch_size=batch_size, shuffle=True, drop_last=True, num_workers=8)
66 | t_loader = torch.utils.data.DataLoader(dataset.Office(t_list_path),
67 | batch_size=batch_size, shuffle=True, drop_last=True, num_workers=8)
68 | val_loader = torch.utils.data.DataLoader(dataset.Office(t_list_path, training=False),
69 | batch_size=1, num_workers=8)
70 |
71 | s_loader_len, t_loader_len = len(s_loader), len(t_loader)
72 |
73 | model = AlexNet(cudable=cuda, n_class=n_class)
74 | if cuda:
75 | model.cuda()
76 |
77 | opt, opt_D = model.get_optimizer(init_lr, lr_mult, lr_mult_D)
78 |
79 | # resume or init
80 | if not resume == '':
81 | pretrain = torch.load(pretrain_path)
82 | model.load_state_dict(pretrain['model'])
83 | opt.load_state_dict(pretrain['opt'])
84 | opt_D.load_state_dict(pretrain['opt_D'])
85 | epoch = pretrain['epoch'] # need change to 0 when DA
86 | else:
87 | model.load_state_dict(utils.load_pth_model(), strict=False)
88 | # model.load_state_dict(utils.load_pretrain_npy(), strict=False)
89 | epoch = 0
90 |
91 | output(' ======= START TRAINING ======= ')
92 | for epoch in range(epoch, 100000):
93 | model.train()
94 | lamb = adaptation_factor(epoch * 1.0 / max_epoch)
95 | current_lr, _ = lr_schedule(opt, epoch, lr_mult), lr_schedule(opt_D, epoch, lr_mult_D)
96 |
97 | if epoch % s_loader_len == 0:
98 | s_loader_epoch = iter(s_loader)
99 | if epoch % t_loader_len == 0:
100 | t_loader_epoch = iter(t_loader)
101 | xs, ys = s_loader_epoch.next()
102 | xt, yt = t_loader_epoch.next()
103 | if cuda:
104 | xs, ys, xt, yt = xs.cuda(), ys.cuda(), xt.cuda(), yt.cuda()
105 |
106 | # forward
107 | s_feature, s_score, s_pred = model.forward(xs)
108 | t_feature, t_score, t_pred = model.forward(xt)
109 | C_loss = model.closs(s_score, ys)
110 | if da:
111 | s_logit, t_logit = model.forward_D(s_feature), model.forward_D(t_feature)
112 |
113 | G_loss, D_loss, semantic_loss = model.adloss(s_logit, t_logit, s_feature, t_feature, ys, t_pred)
114 | Dregloss, Gregloss = model.regloss()
115 | F_loss = C_loss + Gregloss + lamb * G_loss + lamb * semantic_loss
116 | D_loss = D_loss + Dregloss
117 |
118 | opt_D.zero_grad()
119 | D_loss.backward(retain_graph=True)
120 | opt_D.step()
121 | opt.zero_grad()
122 | F_loss.backward(retain_graph=True)
123 | opt.step()
124 | else:
125 | opt.zero_grad()
126 | C_loss.backward()
127 | opt.step()
128 |
129 | if epoch % 10 == 0:
130 | s_pred_label = torch.max(s_score, 1)[1]
131 | s_correct = torch.sum(torch.eq(s_pred_label, ys).float())
132 | s_acc = torch.div(s_correct, ys.size(0))
133 |
134 | output('epoch: {}, lr: {}, lambda: {}'.format(epoch, current_lr, lamb))
135 | if da:
136 | output('correct: {}, C_loss: {}, G_loss:{}, D_loss:{}, Gregloss: {}, Dregloss: {}, semantic_loss: {}, F_loss: {}'.format(
137 | s_correct.item(), C_loss.item(), G_loss.item(), D_loss.item(),
138 | Gregloss.item(), Dregloss.item(), semantic_loss.item(), F_loss.item()))
139 | else:
140 | output('correct: {}, C_loss: {}'.format(s_correct.item(), C_loss.item()))
141 |
142 |
143 | # validation
144 | if epoch % 100 == 0 and epoch != 0:
145 | output(' ======= START VALIDATION ======= ')
146 | model.eval()
147 | v_correct, v_sum = 0, 0
148 | zeros, zeros_classes = torch.zeros(n_class), torch.zeros(n_class)
149 | if cuda:
150 | zeros, zeros_classes = zeros.cuda(), zeros_classes.cuda()
151 | for ind2, (xv, yv) in enumerate(val_loader):
152 | if cuda:
153 | xv, yv = xv.cuda(), yv.cuda()
154 | v_feature, v_score, v_pred = model.forward(xv)
155 | v_pred_label = torch.max(v_score, 1)[1]
156 | v_equal = torch.eq(v_pred_label, yv).float()
157 | zeros = zeros.scatter_add(0, yv, v_equal)
158 | zeros_classes = zeros_classes.scatter_add(0, yv, torch.ones_like(yv, dtype=torch.float))
159 | v_correct += torch.sum(v_equal).item()
160 | v_sum += len(yv)
161 | v_acc = v_correct / v_sum
162 | output('validation: {}, {}'.format(v_correct, v_acc, zeros))
163 | output('class: {}'.format(zeros.tolist()))
164 | output('class: {}'.format(zeros_classes.tolist()))
165 | output('source: {}, target: {}, batch_size: {}, init_lr: {}'.format(s_name, t_name, batch_size, init_lr))
166 | output('lr_mult: {}, lr_mult_D: {}'.format(lr_mult, lr_mult_D))
167 | output(' ======= START TRAINING ======= ')
168 |
169 | # save model
170 | if epoch % 1000 == 0 and epoch != 0:
171 | torch.save({
172 | 'epoch': epoch + 1,
173 | 'model': model.state_dict(),
174 | 'opt': opt.state_dict(),
175 | 'opt_D': opt_D.state_dict()
176 | }, checkpoint_save_path)
177 |
178 | epoch += 1
179 |
--------------------------------------------------------------------------------
/utils.py:
--------------------------------------------------------------------------------
1 | import torch
2 | import torch.nn as nn
3 | import collections
4 | import numpy as np
5 |
6 |
7 | # load pre-trained alexnet model, and return a new Dictionary with matched keys
8 | def load_pretrain_npy():
9 | old_dict = np.load('bvlc_alexnet.npy', encoding='bytes').item()
10 | new_dict = collections.OrderedDict()
11 | for key in old_dict:
12 | if key == 'conv1':
13 | newkey = 'conv.0'
14 | elif key == 'conv2':
15 | newkey = 'conv.4'
16 | elif key == 'conv3':
17 | newkey = 'conv.8'
18 | elif key == 'conv4':
19 | newkey = 'conv.10'
20 | elif key == 'conv5':
21 | newkey = 'conv.12'
22 | elif key == 'fc6':
23 | newkey = 'dense.0'
24 | elif key == 'fc7':
25 | newkey = 'dense.3'
26 | else:
27 | continue
28 | weight = old_dict[key][0]
29 | bias = old_dict[key][1]
30 |
31 | # reverse all dimension for matching, shape==2 is fc, shape==4 is conv
32 | if len(weight.shape) == 2:
33 | weight = np.transpose(weight, (1, 0))
34 | elif len(weight.shape) == 4:
35 | weight = np.transpose(weight, (3, 2, 0, 1))
36 |
37 | # add keys and data
38 | t = torch.tensor(weight)
39 | new_dict[newkey + '.weight'] = t
40 | new_dict[newkey + '.bias'] = torch.Tensor(bias)
41 |
42 | return new_dict
43 |
44 | def load_pth_model():
45 | model_path = '../model/alexnet.pth.tar'
46 | pretrained_model = torch.load(model_path)
47 | return pretrained_model['state_dict']
48 |
49 |
50 | def truncated_normal_(tensor, mean=0, std=0.01):
51 | size = tensor.shape
52 | tmp = tensor.new_empty(size + (4,)).normal_()
53 | valid = (tmp < 2) & (tmp > -2)
54 | ind = valid.max(-1, keepdim=True)[1]
55 | tensor.data.copy_(tmp.gather(-1, ind).squeeze(-1))
56 | tensor.data.mul_(std).add_(mean)
57 |
58 | class LRN(nn.Module):
59 | def __init__(self, local_size=1, alpha=1.0, beta=0.75, ACROSS_CHANNELS=True):
60 | super(LRN, self).__init__()
61 | self.ACROSS_CHANNELS = ACROSS_CHANNELS
62 | if ACROSS_CHANNELS:
63 | self.average = nn.AvgPool3d(kernel_size=(local_size, 1, 1),
64 | stride=1,
65 | padding=(int((local_size - 1.0) / 2), 0, 0))
66 | else:
67 | self.average = nn.AvgPool2d(kernel_size=local_size,
68 | stride=1,
69 | padding=int((local_size - 1.0) / 2))
70 | self.alpha = alpha
71 | self.beta = beta
72 |
73 | def forward(self, x):
74 | if self.ACROSS_CHANNELS:
75 | div = x.pow(2).unsqueeze(1)
76 | div = self.average(div).squeeze(1)
77 | div = div.mul(self.alpha).add(1.0).pow(self.beta)
78 | else:
79 | div = x.pow(2)
80 | div = self.average(div)
81 | div = div.mul(self.alpha).add(1.0).pow(self.beta)
82 | x = x.div(div)
83 | return x
--------------------------------------------------------------------------------