def decoded_output_boxes_class_agnostic(self):
""" Returns: Nx4 """
assert self._bbox_class_agnostic
box_logits = tf.reshape(self.box_logits, [-1, 4])
decoded = decode_bbox_target(box_logits / self.bbox_regression_weights,
self.proposals.boxes)
return decoded
def decoded_output_boxes(self):
""" Returns: N x #class x 4 """
anchors = tf.tile(tf.expand_dims(self.proposals.boxes, 1),
[1, cfg.DATA.NUM_CLASS, 1]) # N x #class x 4
decoded_boxes = decode_bbox_target(
self.box_logits / self.bbox_regression_weights, anchors)
return decoded_boxes
def decode_boxes(self, image_shape2d, rcnn_boxes, rcnn_box_logits, stage_num):
"""
Args:
image_shape2d: h, w
rcnn_boxes (nx4): the proposal boxes
rcnn_box_logits (nx #class x 4):
stage_num:
Returns:
boxes (mx4):
"""
prefix = ''
if stage_num == 2:
prefix = '_2nd'
bbox_reg_weights = cfg.CASCADERCNN.BBOX_REG_WEIGHTS_STAGE2
elif stage_num == 3:
prefix= '3rd'
bbox_reg_weights = cfg.CASCADERCNN.BBOX_REG_WEIGHTS_STAGE3
rcnn_box_logits = rcnn_box_logits[:, 1:, :]
rcnn_box_logits.set_shape([None, cfg.DATA.NUM_CATEGORY, None])
anchors = tf.tile(tf.expand_dims(rcnn_boxes, 1), [1, cfg.DATA.NUM_CATEGORY, 1]) # #proposal x #Cat x 4
decoded_boxes = decode_bbox_target(
rcnn_box_logits /
tf.constant(bbox_reg_weights, dtype=tf.float32), anchors) # #proposal x #Cat x 4
decoded_boxes = clip_boxes(decoded_boxes, image_shape2d, name='fastrcnn_all_boxes'+prefix)
assert decoded_boxes.shape[1] == cfg.DATA.NUM_CATEGORY
decoded_boxes = tf.reshape(decoded_boxes, [-1, 4])
return decoded_boxes
def fastrcnn_inference(self, image_shape2d,
rcnn_boxes, rcnn_label_logits, rcnn_box_logits):
"""
Args:
image_shape2d: h, w
rcnn_boxes (nx4): the proposal boxes
rcnn_label_logits (n):
rcnn_box_logits (nx #class x 4):
Returns:
boxes (mx4):
labels (m): each >= 1
"""
rcnn_box_logits = rcnn_box_logits[:, 1:, :]
rcnn_box_logits.set_shape([None, cfg.DATA.NUM_CATEGORY, None])
label_probs = tf.nn.softmax(rcnn_label_logits, name='fastrcnn_all_probs') # #proposal x #Class
anchors = tf.tile(tf.expand_dims(rcnn_boxes, 1), [1, cfg.DATA.NUM_CATEGORY, 1]) # #proposal x #Cat x 4
decoded_boxes = decode_bbox_target(
rcnn_box_logits /
tf.constant(cfg.FRCNN.BBOX_REG_WEIGHTS, dtype=tf.float32), anchors)
decoded_boxes = clip_boxes(decoded_boxes, image_shape2d, name='fastrcnn_all_boxes')
# indices: Nx2. Each index into (#proposal, #category)
pred_indices, final_probs = fastrcnn_predictions(decoded_boxes, label_probs)
final_probs = tf.identity(final_probs, 'final_probs')
final_boxes = tf.gather_nd(decoded_boxes, pred_indices, name='final_boxes')
final_labels = tf.add(pred_indices[:, 1], 1, name='final_labels')
return final_boxes, final_labels
def fastrcnn_inference(self, image_shape2d,
rcnn_boxes, rcnn_label_logits, rcnn_box_logits):
"""
Args:
image_shape2d: h, w
rcnn_boxes (nx4): the proposal boxes
rcnn_label_logits (n):
rcnn_box_logits (nx #class x 4):
Returns:
boxes (mx4):
labels (m): each >= 1
"""
rcnn_box_logits = rcnn_box_logits[:, 1:, :]
rcnn_box_logits.set_shape([None, cfg.DATA.NUM_CATEGORY, None])
label_probs = tf.nn.softmax(rcnn_label_logits, name='fastrcnn_all_probs') # #proposal x #Class
anchors = tf.tile(tf.expand_dims(rcnn_boxes, 1), [1, cfg.DATA.NUM_CATEGORY, 1]) # #proposal x #Cat x 4
decoded_boxes = decode_bbox_target(
rcnn_box_logits /
tf.constant(cfg.FRCNN.BBOX_REG_WEIGHTS, dtype=tf.float32), anchors)
decoded_boxes = clip_boxes(decoded_boxes, image_shape2d, name='fastrcnn_all_boxes')
# indices: Nx2. Each index into (#proposal, #category)
pred_indices, final_probs = fastrcnn_predictions(decoded_boxes, label_probs)
final_probs = tf.identity(final_probs, 'final_probs')
final_boxes = tf.gather_nd(decoded_boxes, pred_indices, name='final_boxes')
final_labels = tf.add(pred_indices[:, 1], 1, name='final_labels')
return final_boxes, final_labels
def decoded_output_boxes_for_label(self, labels):
assert not self._bbox_class_agnostic
indices = tf.stack(
[tf.range(tf.size(labels, out_type=tf.int64)), labels])
needed_logits = tf.gather_nd(self.box_logits, indices)
decoded = decode_bbox_target(
needed_logits / self.bbox_regression_weights, self.proposals.boxes)
return decoded
def decoded_output_boxes_for_true_label(self):
""" Returns: Nx4 decoded boxes """
indices = tf.stack(
[tf.range(tf.size(self.labels, out_type=tf.int64)), self.labels])
needed_logits = tf.gather_nd(self.box_logits, indices)
decoded = decode_bbox_target(
needed_logits / self.bbox_regression_weights, self.input_boxes)
return decoded
def decoded_output_boxes_class_agnostic(self):
""" Returns: Nx4 """
assert self._bbox_class_agnostic
box_logits = tf.reshape(self.box_logits, [-1, 4])
decoded = decode_bbox_target(
box_logits / self.bbox_regression_weights,
self.proposals.boxes
)
return decoded
def decoded_output_boxes(self):
""" Returns: N x #class x 4 """
anchors = tf.tile(tf.expand_dims(self.proposals.boxes, 1),
[1, cfg.DATA.NUM_CLASS, 1]) # N x #class x 4
decoded_boxes = decode_bbox_target(
self.box_logits / self.bbox_regression_weights,
anchors
)
return decoded_boxes
def decoded_output_boxes_batch(self):
""" Returns: N x #class x 4 """
batch_ids, nobatch_proposal_boxes = tf.split(self.proposal_boxes,
[1, 4], 1)
anchors = tf.tile(tf.expand_dims(nobatch_proposal_boxes, 1),
[1, cfg.DATA.NUM_CLASS, 1]) # N x #class x 4
decoded_boxes = decode_bbox_target(
self.box_logits / self.bbox_regression_weights, anchors)
return decoded_boxes, tf.reshape(batch_ids, [-1])
def decoded_output_boxes_for_label(self, labels):
assert not self._bbox_class_agnostic
indices = tf.stack([
tf.range(tf.size(labels, out_type=tf.int64)),
labels
])
needed_logits = tf.gather_nd(self.box_logits, indices)
decoded = decode_bbox_target(
needed_logits / self.bbox_regression_weights,
self.proposals.boxes
)
return decoded
def fastrcnn_inference_cascade(self, image_shape2d,
rcnn_boxes, rcnn_label_logits, rcnn_box_logits, stage_num):
"""
Args:
image_shape2d: h, w
rcnn_boxes (nx4): the proposal boxes
rcnn_label_logits (n):
rcnn_box_logits (nx #class x 4):
Returns:
boxes (mx4):
labels (m): each >= 1
"""
if stage_num == 1:
bbox_reg_weights = cfg.CASCADERCNN.BBOX_REG_WEIGHTS_STAGE1
elif stage_num == 2:
bbox_reg_weights = cfg.CASCADERCNN.BBOX_REG_WEIGHTS_STAGE2
elif stage_num == 3:
bbox_reg_weights = cfg.CASCADERCNN.BBOX_REG_WEIGHTS_STAGE3
prefix = ''
if stage_num == 1:
prefix = '_1st'
elif stage_num == 2:
prefix = '_2nd'
elif stage_num == 3:
prefix ='_3rd'
rcnn_box_logits = rcnn_box_logits[:, 1:, :]
rcnn_box_logits.set_shape([None, cfg.DATA.NUM_CATEGORY, None])
label_probs = tf.nn.softmax(rcnn_label_logits, name='fastrcnn_all_probs') # #proposal x #Class
anchors = tf.tile(tf.expand_dims(rcnn_boxes, 1), [1, cfg.DATA.NUM_CATEGORY, 1]) # #proposal x #Cat x 4
decoded_boxes = decode_bbox_target(
rcnn_box_logits /
tf.constant(bbox_reg_weights, dtype=tf.float32), anchors)
decoded_boxes = clip_boxes(decoded_boxes, image_shape2d, name='fastrcnn_all_boxes')
# indices: Nx2. Each index into (#proposal, #category)
#TODO add box voting after NMS
if cfg.TEST.BOX_VOTING.ENABLED:
final_boxes, final_probs, pred_indices = fastrcnn_predictions_box_voting(decoded_boxes, label_probs)
final_probs = tf.identity(final_probs, 'final_probs'+prefix)
final_boxes = tf.identity(final_boxes, 'final_boxes'+prefix)
final_labels = tf.add(pred_indices[:, 1], 1, name='final_labels'+prefix)
else:
pred_indices, final_probs = fastrcnn_predictions(decoded_boxes, label_probs)
final_probs = tf.identity(final_probs, 'final_probs'+prefix)
final_boxes = tf.gather_nd(decoded_boxes, pred_indices, name='final_boxes'+prefix)
final_labels = tf.add(pred_indices[:, 1], 1, name='final_labels'+prefix)
return final_boxes, final_labels
def build_graph(self, *inputs):
num_fpn_level = len(cfg.FPN.ANCHOR_STRIDES)
assert len(cfg.RPN.ANCHOR_SIZES) == num_fpn_level
is_training = get_current_tower_context().is_training
image = inputs[0]
input_anchors = inputs[1:1 + 2 * num_fpn_level]
multilevel_anchor_labels = input_anchors[0::2]
multilevel_anchor_boxes = input_anchors[1::2]
gt_boxes, gt_labels = inputs[11], inputs[12]
if cfg.MODE_MASK:
gt_masks = inputs[-1]
image = self.preprocess(image) # 1CHW
image_shape2d = tf.shape(image)[2:] # h,w
c2345 = resnet_fpn_backbone(image, cfg.BACKBONE.RESNET_NUM_BLOCK)
p23456 = fpn_model('fpn', c2345)
# Images are padded for p5, which are too large for p2-p4.
# This seems to have no effect on mAP.
for i, stride in enumerate(cfg.FPN.ANCHOR_STRIDES[:3]):
pi = p23456[i]
target_shape = tf.to_int32(
tf.ceil(tf.to_float(image_shape2d) * (1.0 / stride)))
p23456[i] = tf.slice(pi, [0, 0, 0, 0],
tf.concat([[-1, -1], target_shape], axis=0))
p23456[i].set_shape([1, pi.shape[1], None, None])
# Multi-Level RPN Proposals
multilevel_proposals = []
rpn_loss_collection = []
for lvl in range(num_fpn_level):
rpn_label_logits, rpn_box_logits = rpn_head(
'rpn', p23456[lvl], cfg.FPN.NUM_CHANNEL,
len(cfg.RPN.ANCHOR_RATIOS))
with tf.name_scope('FPN_lvl{}'.format(lvl + 2)):
anchors = tf.constant(get_all_anchors_fpn()[lvl],
name='rpn_anchor_lvl{}'.format(lvl + 2))
anchors, anchor_labels, anchor_boxes = \
self.narrow_to_featuremap(p23456[lvl], anchors,
multilevel_anchor_labels[lvl],
multilevel_anchor_boxes[lvl])
anchor_boxes_encoded = encode_bbox_target(
anchor_boxes, anchors)
pred_boxes_decoded = decode_bbox_target(
rpn_box_logits, anchors)
proposal_boxes, proposal_scores = generate_rpn_proposals(
tf.reshape(pred_boxes_decoded, [-1, 4]),
tf.reshape(rpn_label_logits, [-1]), image_shape2d,
cfg.RPN.TRAIN_FPN_NMS_TOPK
if is_training else cfg.RPN.TEST_FPN_NMS_TOPK)
multilevel_proposals.append((proposal_boxes, proposal_scores))
if is_training:
label_loss, box_loss = rpn_losses(anchor_labels,
anchor_boxes_encoded,
rpn_label_logits,
rpn_box_logits)
rpn_loss_collection.extend([label_loss, box_loss])
# Merge proposals from multi levels, pick top K
proposal_boxes = tf.concat([x[0] for x in multilevel_proposals],
axis=0) # nx4
proposal_scores = tf.concat([x[1] for x in multilevel_proposals],
axis=0) # n
proposal_topk = tf.minimum(
tf.size(proposal_scores), cfg.RPN.TRAIN_FPN_NMS_TOPK
if is_training else cfg.RPN.TEST_FPN_NMS_TOPK)
proposal_scores, topk_indices = tf.nn.top_k(proposal_scores,
k=proposal_topk,
sorted=False)
proposal_boxes = tf.gather(proposal_boxes, topk_indices)
if is_training:
rcnn_boxes, rcnn_labels, fg_inds_wrt_gt = sample_fast_rcnn_targets(
proposal_boxes, gt_boxes, gt_labels)
else:
# The boxes to be used to crop RoIs.
rcnn_boxes = proposal_boxes
roi_feature_fastrcnn = multilevel_roi_align(p23456[:4], rcnn_boxes, 7)
fastrcnn_head_func = getattr(model, cfg.FPN.FRCNN_HEAD_FUNC)
fastrcnn_label_logits, fastrcnn_box_logits = fastrcnn_head_func(
'fastrcnn', roi_feature_fastrcnn, cfg.DATA.NUM_CLASS)
if is_training:
# rpn loss is already defined above
with tf.name_scope('rpn_losses'):
rpn_total_label_loss = tf.add_n(rpn_loss_collection[::2],
name='label_loss')
rpn_total_box_loss = tf.add_n(rpn_loss_collection[1::2],
name='box_loss')
add_moving_summary(rpn_total_box_loss, rpn_total_label_loss)
# fastrcnn loss:
matched_gt_boxes = tf.gather(gt_boxes, fg_inds_wrt_gt)
fg_inds_wrt_sample = tf.reshape(tf.where(rcnn_labels > 0),
[-1]) # fg inds w.r.t all samples
fg_sampled_boxes = tf.gather(rcnn_boxes, fg_inds_wrt_sample)
fg_fastrcnn_box_logits = tf.gather(fastrcnn_box_logits,
fg_inds_wrt_sample)
fastrcnn_label_loss, fastrcnn_box_loss = self.fastrcnn_training(
image, rcnn_labels, fg_sampled_boxes, matched_gt_boxes,
fastrcnn_label_logits, fg_fastrcnn_box_logits)
if cfg.MODE_MASK:
# maskrcnn loss
fg_labels = tf.gather(rcnn_labels, fg_inds_wrt_sample)
roi_feature_maskrcnn = multilevel_roi_align(
p23456[:4], fg_sampled_boxes, 14)
mask_logits = maskrcnn_upXconv_head('maskrcnn',
roi_feature_maskrcnn,
cfg.DATA.NUM_CATEGORY,
4) # #fg x #cat x 28 x 28
target_masks_for_fg = crop_and_resize(
tf.expand_dims(gt_masks, 1),
fg_sampled_boxes,
fg_inds_wrt_gt,
28,
pad_border=False) # fg x 1x28x28
target_masks_for_fg = tf.squeeze(target_masks_for_fg, 1,
'sampled_fg_mask_targets')
mrcnn_loss = maskrcnn_loss(mask_logits, fg_labels,
target_masks_for_fg)
else:
mrcnn_loss = 0.0
wd_cost = regularize_cost(
'(?:group1|group2|group3|rpn|fpn|fastrcnn|maskrcnn)/.*W',
l2_regularizer(cfg.TRAIN.WEIGHT_DECAY),
name='wd_cost')
total_cost = tf.add_n(
rpn_loss_collection +
[fastrcnn_label_loss, fastrcnn_box_loss, mrcnn_loss, wd_cost],
'total_cost')
add_moving_summary(total_cost, wd_cost)
return total_cost
else:
final_boxes, final_labels = self.fastrcnn_inference(
image_shape2d, rcnn_boxes, fastrcnn_label_logits,
fastrcnn_box_logits)
if cfg.MODE_MASK:
# Cascade inference needs roi transform with refined boxes.
roi_feature_maskrcnn = multilevel_roi_align(
p23456[:4], final_boxes, 14)
mask_logits = maskrcnn_upXconv_head('maskrcnn',
roi_feature_maskrcnn,
cfg.DATA.NUM_CATEGORY,
4) # #fg x #cat x 28 x 28
indices = tf.stack([
tf.range(tf.size(final_labels)),
tf.to_int32(final_labels) - 1
],
axis=1)
final_mask_logits = tf.gather_nd(mask_logits,
indices) # #resultx28x28
tf.sigmoid(final_mask_logits, name='final_masks')
def build_graph(self, *inputs):
is_training = get_current_tower_context().is_training
if cfg.MODE_MASK:
image, anchor_labels, anchor_boxes, gt_boxes, gt_labels, gt_masks = inputs
else:
image, anchor_labels, anchor_boxes, gt_boxes, gt_labels = inputs
image = self.preprocess(image) # 1CHW
featuremap = resnet_c4_backbone(image,
cfg.BACKBONE.RESNET_NUM_BLOCK[:3])
rpn_label_logits, rpn_box_logits = rpn_head('rpn', featuremap, 1024,
cfg.RPN.NUM_ANCHOR)
fm_anchors, anchor_labels, anchor_boxes = self.narrow_to_featuremap(
featuremap, get_all_anchors(), anchor_labels, anchor_boxes)
anchor_boxes_encoded = encode_bbox_target(anchor_boxes, fm_anchors)
image_shape2d = tf.shape(image)[2:] # h,w
pred_boxes_decoded = decode_bbox_target(
rpn_box_logits, fm_anchors) # fHxfWxNAx4, floatbox
proposal_boxes, proposal_scores = generate_rpn_proposals(
tf.reshape(pred_boxes_decoded, [-1, 4]),
tf.reshape(rpn_label_logits,
[-1]), image_shape2d, cfg.RPN.TRAIN_PRE_NMS_TOPK
if is_training else cfg.RPN.TEST_PRE_NMS_TOPK,
cfg.RPN.TRAIN_POST_NMS_TOPK
if is_training else cfg.RPN.TEST_POST_NMS_TOPK)
if is_training:
# sample proposal boxes in training
rcnn_boxes, rcnn_labels, fg_inds_wrt_gt = sample_fast_rcnn_targets(
proposal_boxes, gt_boxes, gt_labels)
else:
# The boxes to be used to crop RoIs.
# Use all proposal boxes in inference
rcnn_boxes = proposal_boxes
boxes_on_featuremap = rcnn_boxes * (1.0 / cfg.RPN.ANCHOR_STRIDE)
roi_resized = roi_align(featuremap, boxes_on_featuremap, 14)
feature_fastrcnn = resnet_conv5(
roi_resized, cfg.BACKBONE.RESNET_NUM_BLOCK[-1]) # nxcx7x7
# Keep C5 feature to be shared with mask branch
feature_gap = GlobalAvgPooling('gap',
feature_fastrcnn,
data_format='channels_first')
fastrcnn_label_logits, fastrcnn_box_logits = fastrcnn_outputs(
'fastrcnn', feature_gap, cfg.DATA.NUM_CLASS)
if is_training:
# rpn loss
rpn_label_loss, rpn_box_loss = rpn_losses(anchor_labels,
anchor_boxes_encoded,
rpn_label_logits,
rpn_box_logits)
# fastrcnn loss
matched_gt_boxes = tf.gather(gt_boxes, fg_inds_wrt_gt)
fg_inds_wrt_sample = tf.reshape(tf.where(rcnn_labels > 0),
[-1]) # fg inds w.r.t all samples
fg_sampled_boxes = tf.gather(rcnn_boxes, fg_inds_wrt_sample)
fg_fastrcnn_box_logits = tf.gather(fastrcnn_box_logits,
fg_inds_wrt_sample)
fastrcnn_label_loss, fastrcnn_box_loss = self.fastrcnn_training(
image, rcnn_labels, fg_sampled_boxes, matched_gt_boxes,
fastrcnn_label_logits, fg_fastrcnn_box_logits)
if cfg.MODE_MASK:
# maskrcnn loss
fg_labels = tf.gather(rcnn_labels, fg_inds_wrt_sample)
# In training, mask branch shares the same C5 feature.
fg_feature = tf.gather(feature_fastrcnn, fg_inds_wrt_sample)
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', fg_feature, cfg.DATA.NUM_CATEGORY,
num_convs=0) # #fg x #cat x 14x14
target_masks_for_fg = crop_and_resize(
tf.expand_dims(gt_masks, 1),
fg_sampled_boxes,
fg_inds_wrt_gt,
14,
pad_border=False) # nfg x 1x14x14
target_masks_for_fg = tf.squeeze(target_masks_for_fg, 1,
'sampled_fg_mask_targets')
mrcnn_loss = maskrcnn_loss(mask_logits, fg_labels,
target_masks_for_fg)
else:
mrcnn_loss = 0.0
wd_cost = regularize_cost(
'(?:group1|group2|group3|rpn|fastrcnn|maskrcnn)/.*W',
l2_regularizer(cfg.TRAIN.WEIGHT_DECAY),
name='wd_cost')
total_cost = tf.add_n([
rpn_label_loss, rpn_box_loss, fastrcnn_label_loss,
fastrcnn_box_loss, mrcnn_loss, wd_cost
], 'total_cost')
add_moving_summary(total_cost, wd_cost)
return total_cost
else:
final_boxes, final_labels = self.fastrcnn_inference(
image_shape2d, rcnn_boxes, fastrcnn_label_logits,
fastrcnn_box_logits)
if cfg.MODE_MASK:
roi_resized = roi_align(
featuremap, final_boxes * (1.0 / cfg.RPN.ANCHOR_STRIDE),
14)
feature_maskrcnn = resnet_conv5(
roi_resized, cfg.BACKBONE.RESNET_NUM_BLOCK[-1])
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', feature_maskrcnn, cfg.DATA.NUM_CATEGORY,
0) # #result x #cat x 14x14
indices = tf.stack([
tf.range(tf.size(final_labels)),
tf.to_int32(final_labels) - 1
],
axis=1)
final_mask_logits = tf.gather_nd(mask_logits,
indices) # #resultx14x14
tf.sigmoid(final_mask_logits, name='final_masks')
def build_graph(self, *inputs):
num_fpn_level = len(cfg.FPN.ANCHOR_STRIDES)
assert len(cfg.RPN.ANCHOR_SIZES) == num_fpn_level
is_training = get_current_tower_context().is_training
image = inputs[0]
input_anchors = inputs[1: 1 + 2 * num_fpn_level]
multilevel_anchor_labels = input_anchors[0::2]
multilevel_anchor_boxes = input_anchors[1::2]
gt_boxes, gt_labels = inputs[11], inputs[12]
if cfg.MODE_MASK:
gt_masks = inputs[-1]
image = self.preprocess(image) # 1CHW
image_shape2d = tf.shape(image)[2:] # h,w
c2345 = resnet_fpn_backbone(image, cfg.BACKBONE.RESNET_NUM_BLOCK)
p23456 = fpn_model('fpn', c2345)
# Images are padded for p5, which are too large for p2-p4.
# This seems to have no effect on mAP.
for i, stride in enumerate(cfg.FPN.ANCHOR_STRIDES[:3]):
pi = p23456[i]
target_shape = tf.to_int32(tf.ceil(tf.to_float(image_shape2d) * (1.0 / stride)))
p23456[i] = tf.slice(pi, [0, 0, 0, 0],
tf.concat([[-1, -1], target_shape], axis=0))
p23456[i].set_shape([1, pi.shape[1], None, None])
# Multi-Level RPN Proposals
multilevel_proposals = []
rpn_loss_collection = []
for lvl in range(num_fpn_level):
rpn_label_logits, rpn_box_logits = rpn_head(
'rpn', p23456[lvl], cfg.FPN.NUM_CHANNEL, len(cfg.RPN.ANCHOR_RATIOS))
with tf.name_scope('FPN_lvl{}'.format(lvl + 2)):
anchors = tf.constant(get_all_anchors_fpn()[lvl], name='rpn_anchor_lvl{}'.format(lvl + 2))
anchors, anchor_labels, anchor_boxes = \
self.narrow_to_featuremap(p23456[lvl], anchors,
multilevel_anchor_labels[lvl],
multilevel_anchor_boxes[lvl])
anchor_boxes_encoded = encode_bbox_target(anchor_boxes, anchors)
pred_boxes_decoded = decode_bbox_target(rpn_box_logits, anchors)
proposal_boxes, proposal_scores = generate_rpn_proposals(
tf.reshape(pred_boxes_decoded, [-1, 4]),
tf.reshape(rpn_label_logits, [-1]),
image_shape2d,
cfg.RPN.TRAIN_FPN_NMS_TOPK if is_training else cfg.RPN.TEST_FPN_NMS_TOPK)
multilevel_proposals.append((proposal_boxes, proposal_scores))
if is_training:
label_loss, box_loss = rpn_losses(
anchor_labels, anchor_boxes_encoded,
rpn_label_logits, rpn_box_logits)
rpn_loss_collection.extend([label_loss, box_loss])
# Merge proposals from multi levels, pick top K
proposal_boxes = tf.concat([x[0] for x in multilevel_proposals], axis=0) # nx4
proposal_scores = tf.concat([x[1] for x in multilevel_proposals], axis=0) # n
proposal_topk = tf.minimum(tf.size(proposal_scores),
cfg.RPN.TRAIN_FPN_NMS_TOPK if is_training else cfg.RPN.TEST_FPN_NMS_TOPK)
proposal_scores, topk_indices = tf.nn.top_k(proposal_scores, k=proposal_topk, sorted=False)
proposal_boxes = tf.gather(proposal_boxes, topk_indices)
if is_training:
rcnn_boxes, rcnn_labels, fg_inds_wrt_gt = sample_fast_rcnn_targets(
proposal_boxes, gt_boxes, gt_labels)
else:
# The boxes to be used to crop RoIs.
rcnn_boxes = proposal_boxes
roi_feature_fastrcnn = multilevel_roi_align(p23456[:4], rcnn_boxes, 7)
fastrcnn_head_func = getattr(model, cfg.FPN.FRCNN_HEAD_FUNC)
fastrcnn_label_logits, fastrcnn_box_logits = fastrcnn_head_func(
'fastrcnn', roi_feature_fastrcnn, cfg.DATA.NUM_CLASS)
if is_training:
# rpn loss is already defined above
with tf.name_scope('rpn_losses'):
rpn_total_label_loss = tf.add_n(rpn_loss_collection[::2], name='label_loss')
rpn_total_box_loss = tf.add_n(rpn_loss_collection[1::2], name='box_loss')
add_moving_summary(rpn_total_box_loss, rpn_total_label_loss)
# fastrcnn loss:
matched_gt_boxes = tf.gather(gt_boxes, fg_inds_wrt_gt)
fg_inds_wrt_sample = tf.reshape(tf.where(rcnn_labels > 0), [-1]) # fg inds w.r.t all samples
fg_sampled_boxes = tf.gather(rcnn_boxes, fg_inds_wrt_sample)
fg_fastrcnn_box_logits = tf.gather(fastrcnn_box_logits, fg_inds_wrt_sample)
fastrcnn_label_loss, fastrcnn_box_loss = self.fastrcnn_training(
image, rcnn_labels, fg_sampled_boxes,
matched_gt_boxes, fastrcnn_label_logits, fg_fastrcnn_box_logits)
if cfg.MODE_MASK:
# maskrcnn loss
fg_labels = tf.gather(rcnn_labels, fg_inds_wrt_sample)
roi_feature_maskrcnn = multilevel_roi_align(
p23456[:4], fg_sampled_boxes, 14)
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', roi_feature_maskrcnn, cfg.DATA.NUM_CATEGORY, 4) # #fg x #cat x 28 x 28
target_masks_for_fg = crop_and_resize(
tf.expand_dims(gt_masks, 1),
fg_sampled_boxes,
fg_inds_wrt_gt, 28,
pad_border=False) # fg x 1x28x28
target_masks_for_fg = tf.squeeze(target_masks_for_fg, 1, 'sampled_fg_mask_targets')
mrcnn_loss = maskrcnn_loss(mask_logits, fg_labels, target_masks_for_fg)
else:
mrcnn_loss = 0.0
wd_cost = regularize_cost(
'(?:group1|group2|group3|rpn|fpn|fastrcnn|maskrcnn)/.*W',
l2_regularizer(cfg.TRAIN.WEIGHT_DECAY), name='wd_cost')
total_cost = tf.add_n(rpn_loss_collection + [
fastrcnn_label_loss, fastrcnn_box_loss,
mrcnn_loss, wd_cost], 'total_cost')
add_moving_summary(total_cost, wd_cost)
return total_cost
else:
final_boxes, final_labels = self.fastrcnn_inference(
image_shape2d, rcnn_boxes, fastrcnn_label_logits, fastrcnn_box_logits)
if cfg.MODE_MASK:
# Cascade inference needs roi transform with refined boxes.
roi_feature_maskrcnn = multilevel_roi_align(p23456[:4], final_boxes, 14)
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', roi_feature_maskrcnn, cfg.DATA.NUM_CATEGORY, 4) # #fg x #cat x 28 x 28
indices = tf.stack([tf.range(tf.size(final_labels)), tf.to_int32(final_labels) - 1], axis=1)
final_mask_logits = tf.gather_nd(mask_logits, indices) # #resultx28x28
tf.sigmoid(final_mask_logits, name='final_masks')
def build_graph(self, *inputs):
is_training = get_current_tower_context().is_training
if cfg.MODE_MASK:
image, anchor_labels, anchor_boxes, gt_boxes, gt_labels, gt_masks = inputs
else:
image, anchor_labels, anchor_boxes, gt_boxes, gt_labels = inputs
image = self.preprocess(image) # 1CHW
featuremap = resnet_c4_backbone(image, cfg.BACKBONE.RESNET_NUM_BLOCK[:3])
rpn_label_logits, rpn_box_logits = rpn_head('rpn', featuremap, 1024, cfg.RPN.NUM_ANCHOR)
fm_anchors, anchor_labels, anchor_boxes = self.narrow_to_featuremap(
featuremap, get_all_anchors(), anchor_labels, anchor_boxes)
anchor_boxes_encoded = encode_bbox_target(anchor_boxes, fm_anchors)
image_shape2d = tf.shape(image)[2:] # h,w
pred_boxes_decoded = decode_bbox_target(rpn_box_logits, fm_anchors) # fHxfWxNAx4, floatbox
proposal_boxes, proposal_scores = generate_rpn_proposals(
tf.reshape(pred_boxes_decoded, [-1, 4]),
tf.reshape(rpn_label_logits, [-1]),
image_shape2d,
cfg.RPN.TRAIN_PRE_NMS_TOPK if is_training else cfg.RPN.TEST_PRE_NMS_TOPK,
cfg.RPN.TRAIN_POST_NMS_TOPK if is_training else cfg.RPN.TEST_POST_NMS_TOPK)
if is_training:
# sample proposal boxes in training
rcnn_boxes, rcnn_labels, fg_inds_wrt_gt = sample_fast_rcnn_targets(
proposal_boxes, gt_boxes, gt_labels)
else:
# The boxes to be used to crop RoIs.
# Use all proposal boxes in inference
rcnn_boxes = proposal_boxes
boxes_on_featuremap = rcnn_boxes * (1.0 / cfg.RPN.ANCHOR_STRIDE)
roi_resized = roi_align(featuremap, boxes_on_featuremap, 14)
feature_fastrcnn = resnet_conv5(roi_resized, cfg.BACKBONE.RESNET_NUM_BLOCK[-1]) # nxcx7x7
# Keep C5 feature to be shared with mask branch
feature_gap = GlobalAvgPooling('gap', feature_fastrcnn, data_format='channels_first')
fastrcnn_label_logits, fastrcnn_box_logits = fastrcnn_outputs('fastrcnn', feature_gap, cfg.DATA.NUM_CLASS)
if is_training:
# rpn loss
rpn_label_loss, rpn_box_loss = rpn_losses(
anchor_labels, anchor_boxes_encoded, rpn_label_logits, rpn_box_logits)
# fastrcnn loss
matched_gt_boxes = tf.gather(gt_boxes, fg_inds_wrt_gt)
fg_inds_wrt_sample = tf.reshape(tf.where(rcnn_labels > 0), [-1]) # fg inds w.r.t all samples
fg_sampled_boxes = tf.gather(rcnn_boxes, fg_inds_wrt_sample)
fg_fastrcnn_box_logits = tf.gather(fastrcnn_box_logits, fg_inds_wrt_sample)
fastrcnn_label_loss, fastrcnn_box_loss = self.fastrcnn_training(
image, rcnn_labels, fg_sampled_boxes,
matched_gt_boxes, fastrcnn_label_logits, fg_fastrcnn_box_logits)
if cfg.MODE_MASK:
# maskrcnn loss
fg_labels = tf.gather(rcnn_labels, fg_inds_wrt_sample)
# In training, mask branch shares the same C5 feature.
fg_feature = tf.gather(feature_fastrcnn, fg_inds_wrt_sample)
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', fg_feature, cfg.DATA.NUM_CATEGORY, num_convs=0) # #fg x #cat x 14x14
target_masks_for_fg = crop_and_resize(
tf.expand_dims(gt_masks, 1),
fg_sampled_boxes,
fg_inds_wrt_gt, 14,
pad_border=False) # nfg x 1x14x14
target_masks_for_fg = tf.squeeze(target_masks_for_fg, 1, 'sampled_fg_mask_targets')
mrcnn_loss = maskrcnn_loss(mask_logits, fg_labels, target_masks_for_fg)
else:
mrcnn_loss = 0.0
wd_cost = regularize_cost(
'(?:group1|group2|group3|rpn|fastrcnn|maskrcnn)/.*W',
l2_regularizer(cfg.TRAIN.WEIGHT_DECAY), name='wd_cost')
total_cost = tf.add_n([
rpn_label_loss, rpn_box_loss,
fastrcnn_label_loss, fastrcnn_box_loss,
mrcnn_loss,
wd_cost], 'total_cost')
add_moving_summary(total_cost, wd_cost)
return total_cost
else:
final_boxes, final_labels = self.fastrcnn_inference(
image_shape2d, rcnn_boxes, fastrcnn_label_logits, fastrcnn_box_logits)
if cfg.MODE_MASK:
roi_resized = roi_align(featuremap, final_boxes * (1.0 / cfg.RPN.ANCHOR_STRIDE), 14)
feature_maskrcnn = resnet_conv5(roi_resized, cfg.BACKBONE.RESNET_NUM_BLOCK[-1])
mask_logits = maskrcnn_upXconv_head(
'maskrcnn', feature_maskrcnn, cfg.DATA.NUM_CATEGORY, 0) # #result x #cat x 14x14
indices = tf.stack([tf.range(tf.size(final_labels)), tf.to_int32(final_labels) - 1], axis=1)
final_mask_logits = tf.gather_nd(mask_logits, indices) # #resultx14x14
tf.sigmoid(final_mask_logits, name='final_masks')
