-
Notifications
You must be signed in to change notification settings - Fork 37
/
faster_rcnn.py
247 lines (191 loc) · 10.4 KB
/
faster_rcnn.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
import random
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.autograd import Variable
import torchvision.models as models
from torch.autograd import Variable
import numpy as np
from model.utils.config import cfg
from model.rpn.rpn import _RPN
from model.roi_pooling.modules.roi_pool import _RoIPooling
from model.roi_crop.modules.roi_crop import _RoICrop
from model.roi_align.modules.roi_align import RoIAlignAvg
from model.rpn.proposal_target_layer_cascade import _ProposalTargetLayer
from model.da_faster_rcnn.DA import _ImageDA
from model.da_faster_rcnn.DA import _InstanceDA
import time
import pdb
from model.utils.net_utils import _smooth_l1_loss, _crop_pool_layer, _affine_grid_gen, _affine_theta
class _fasterRCNN(nn.Module):
""" faster RCNN """
def __init__(self, classes, class_agnostic):
super(_fasterRCNN, self).__init__()
self.classes = classes
self.n_classes = len(classes)
self.class_agnostic = class_agnostic
# loss
self.RCNN_loss_cls = 0
self.RCNN_loss_bbox = 0
# define rpn
self.RCNN_rpn = _RPN(self.dout_base_model)
self.RCNN_proposal_target = _ProposalTargetLayer(self.n_classes)
self.RCNN_roi_pool = _RoIPooling(cfg.POOLING_SIZE, cfg.POOLING_SIZE, 1.0/16.0)
self.RCNN_roi_align = RoIAlignAvg(cfg.POOLING_SIZE, cfg.POOLING_SIZE, 1.0/16.0)
self.grid_size = cfg.POOLING_SIZE * 2 if cfg.CROP_RESIZE_WITH_MAX_POOL else cfg.POOLING_SIZE
self.RCNN_roi_crop = _RoICrop()
self.RCNN_imageDA = _ImageDA(self.dout_base_model)
self.RCNN_instanceDA = _InstanceDA()
self.consistency_loss = torch.nn.MSELoss(size_average=False)
def forward(self, im_data, im_info, gt_boxes, num_boxes, need_backprop,
tgt_im_data, tgt_im_info, tgt_gt_boxes, tgt_num_boxes, tgt_need_backprop):
assert need_backprop.detach()==1 and tgt_need_backprop.detach()==0
batch_size = im_data.size(0)
im_info = im_info.data #(size1,size2, image ratio(new image / source image) )
gt_boxes = gt_boxes.data
num_boxes = num_boxes.data
need_backprop=need_backprop.data
# feed image data to base model to obtain base feature map
base_feat = self.RCNN_base(im_data)
# feed base feature map tp RPN to obtain rois
self.RCNN_rpn.train()
rois, rpn_loss_cls, rpn_loss_bbox = self.RCNN_rpn(base_feat, im_info, gt_boxes, num_boxes)
# if it is training phrase, then use ground trubut bboxes for refining
if self.training:
roi_data = self.RCNN_proposal_target(rois, gt_boxes, num_boxes)
rois, rois_label, rois_target, rois_inside_ws, rois_outside_ws = roi_data
rois_label = Variable(rois_label.view(-1).long())
rois_target = Variable(rois_target.view(-1, rois_target.size(2)))
rois_inside_ws = Variable(rois_inside_ws.view(-1, rois_inside_ws.size(2)))
rois_outside_ws = Variable(rois_outside_ws.view(-1, rois_outside_ws.size(2)))
else:
rois_label = None
rois_target = None
rois_inside_ws = None
rois_outside_ws = None
rpn_loss_cls = 0
rpn_loss_bbox = 0
rois = Variable(rois)
# do roi pooling based on predicted rois
if cfg.POOLING_MODE == 'crop':
# pdb.set_trace()
# pooled_feat_anchor = _crop_pool_layer(base_feat, rois.view(-1, 5))
grid_xy = _affine_grid_gen(rois.view(-1, 5), base_feat.size()[2:], self.grid_size)
grid_yx = torch.stack([grid_xy.data[:,:,:,1], grid_xy.data[:,:,:,0]], 3).contiguous()
pooled_feat = self.RCNN_roi_crop(base_feat, Variable(grid_yx).detach())
if cfg.CROP_RESIZE_WITH_MAX_POOL:
pooled_feat = F.max_pool2d(pooled_feat, 2, 2)
elif cfg.POOLING_MODE == 'align':
pooled_feat = self.RCNN_roi_align(base_feat, rois.view(-1, 5))
elif cfg.POOLING_MODE == 'pool':
pooled_feat = self.RCNN_roi_pool(base_feat, rois.view(-1,5))
# feed pooled features to top model
pooled_feat = self._head_to_tail(pooled_feat)
# compute bbox offset
bbox_pred = self.RCNN_bbox_pred(pooled_feat)
if self.training and not self.class_agnostic:
# select the corresponding columns according to roi labels
bbox_pred_view = bbox_pred.view(bbox_pred.size(0), int(bbox_pred.size(1) / 4), 4)
bbox_pred_select = torch.gather(bbox_pred_view, 1, rois_label.view(rois_label.size(0), 1, 1).expand(rois_label.size(0), 1, 4))
bbox_pred = bbox_pred_select.squeeze(1)
# compute object classification probability
cls_score = self.RCNN_cls_score(pooled_feat)
cls_prob = F.softmax(cls_score, 1)
RCNN_loss_cls = 0
RCNN_loss_bbox = 0
if self.training:
# classification loss
RCNN_loss_cls = F.cross_entropy(cls_score, rois_label)
# bounding box regression L1 loss
RCNN_loss_bbox = _smooth_l1_loss(bbox_pred, rois_target, rois_inside_ws, rois_outside_ws)
cls_prob = cls_prob.view(batch_size, rois.size(1), -1)
bbox_pred = bbox_pred.view(batch_size, rois.size(1), -1)
""" =================== for target =========================="""
tgt_batch_size = tgt_im_data.size(0)
tgt_im_info = tgt_im_info.data # (size1,size2, image ratio(new image / source image) )
tgt_gt_boxes = tgt_gt_boxes.data
tgt_num_boxes = tgt_num_boxes.data
tgt_need_backprop = tgt_need_backprop.data
# feed image data to base model to obtain base feature map
tgt_base_feat = self.RCNN_base(tgt_im_data)
# feed base feature map tp RPN to obtain rois
self.RCNN_rpn.eval()
tgt_rois, tgt_rpn_loss_cls, tgt_rpn_loss_bbox = \
self.RCNN_rpn(tgt_base_feat, tgt_im_info, tgt_gt_boxes, tgt_num_boxes)
# if it is training phrase, then use ground trubut bboxes for refining
tgt_rois_label = None
tgt_rois_target = None
tgt_rois_inside_ws = None
tgt_rois_outside_ws = None
tgt_rpn_loss_cls = 0
tgt_rpn_loss_bbox = 0
tgt_rois = Variable(tgt_rois)
# do roi pooling based on predicted rois
if cfg.POOLING_MODE == 'crop':
# pdb.set_trace()
# pooled_feat_anchor = _crop_pool_layer(base_feat, rois.view(-1, 5))
tgt_grid_xy = _affine_grid_gen(tgt_rois.view(-1, 5), tgt_base_feat.size()[2:], self.grid_size)
tgt_grid_yx = torch.stack([tgt_grid_xy.data[:, :, :, 1], tgt_grid_xy.data[:, :, :, 0]], 3).contiguous()
tgt_pooled_feat = self.RCNN_roi_crop(tgt_base_feat, Variable(tgt_grid_yx).detach())
if cfg.CROP_RESIZE_WITH_MAX_POOL:
tgt_pooled_feat = F.max_pool2d(tgt_pooled_feat, 2, 2)
elif cfg.POOLING_MODE == 'align':
tgt_pooled_feat = self.RCNN_roi_align(tgt_base_feat, tgt_rois.view(-1, 5))
elif cfg.POOLING_MODE == 'pool':
tgt_pooled_feat = self.RCNN_roi_pool(tgt_base_feat, tgt_rois.view(-1, 5))
# feed pooled features to top model
tgt_pooled_feat = self._head_to_tail(tgt_pooled_feat)
""" DA loss """
# DA LOSS
DA_img_loss_cls = 0
DA_ins_loss_cls = 0
tgt_DA_img_loss_cls = 0
tgt_DA_ins_loss_cls = 0
base_score, base_label = self.RCNN_imageDA(base_feat, need_backprop)
# Image DA
base_prob = F.log_softmax(base_score, dim=1)
DA_img_loss_cls = F.nll_loss(base_prob, base_label)
instance_sigmoid, same_size_label = self.RCNN_instanceDA(pooled_feat, need_backprop)
instance_loss = nn.BCELoss()
DA_ins_loss_cls = instance_loss(instance_sigmoid, same_size_label)
#consistency_prob = torch.max(F.softmax(base_score, dim=1),dim=1)[0]
consistency_prob = F.softmax(base_score, dim=1)[:,1,:,:]
consistency_prob=torch.mean(consistency_prob)
consistency_prob=consistency_prob.repeat(instance_sigmoid.size())
DA_cst_loss=self.consistency_loss(instance_sigmoid,consistency_prob.detach())
""" ************** taget loss **************** """
tgt_base_score, tgt_base_label = \
self.RCNN_imageDA(tgt_base_feat, tgt_need_backprop)
# Image DA
tgt_base_prob = F.log_softmax(tgt_base_score, dim=1)
tgt_DA_img_loss_cls = F.nll_loss(tgt_base_prob, tgt_base_label)
tgt_instance_sigmoid, tgt_same_size_label = \
self.RCNN_instanceDA(tgt_pooled_feat, tgt_need_backprop)
tgt_instance_loss = nn.BCELoss()
tgt_DA_ins_loss_cls = \
tgt_instance_loss(tgt_instance_sigmoid, tgt_same_size_label)
tgt_consistency_prob = F.softmax(tgt_base_score, dim=1)[:, 0, :, :]
tgt_consistency_prob = torch.mean(tgt_consistency_prob)
tgt_consistency_prob = tgt_consistency_prob.repeat(tgt_instance_sigmoid.size())
tgt_DA_cst_loss = self.consistency_loss(tgt_instance_sigmoid, tgt_consistency_prob.detach())
return rois, cls_prob, bbox_pred, rpn_loss_cls, rpn_loss_bbox, RCNN_loss_cls, RCNN_loss_bbox, rois_label,\
DA_img_loss_cls,DA_ins_loss_cls,tgt_DA_img_loss_cls,tgt_DA_ins_loss_cls,DA_cst_loss,tgt_DA_cst_loss
def _init_weights(self):
def normal_init(m, mean, stddev, truncated=False):
"""
weight initalizer: truncated normal and random normal.
"""
# x is a parameter
if truncated:
m.weight.data.normal_().fmod_(2).mul_(stddev).add_(mean) # not a perfect approximation
else:
m.weight.data.normal_(mean, stddev)
m.bias.data.zero_()
normal_init(self.RCNN_rpn.RPN_Conv, 0, 0.01, cfg.TRAIN.TRUNCATED)
normal_init(self.RCNN_rpn.RPN_cls_score, 0, 0.01, cfg.TRAIN.TRUNCATED)
normal_init(self.RCNN_rpn.RPN_bbox_pred, 0, 0.01, cfg.TRAIN.TRUNCATED)
normal_init(self.RCNN_cls_score, 0, 0.01, cfg.TRAIN.TRUNCATED)
normal_init(self.RCNN_bbox_pred, 0, 0.001, cfg.TRAIN.TRUNCATED)
def create_architecture(self):
self._init_modules()
self._init_weights()