def __init__(self, RelDN, box_roi_pool, box_head, box_predictor, # Faster R-CNN training fg_iou_thresh, bg_iou_thresh, batch_size_per_image, positive_fraction, bbox_reg_weights, # Faster R-CNN inference score_thresh, nms_thresh, detections_per_img, # Mask mask_roi_pool=None, mask_head=None, mask_predictor=None, keypoint_roi_pool=None, keypoint_head=None, keypoint_predictor=None, ): super(RoIHeads, self).__init__() batch_size_per_image_so = 64 positive_fraction_so = 0.5 self.box_similarity = box_ops.box_iou self.RelDN = RelDN # assign ground-truth boxes for each proposal self.proposal_matcher = det_utils.Matcher( fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False) self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler( batch_size_per_image, positive_fraction) self.fg_bg_sampler_so = det_utils.BalancedPositiveNegativeSampler( batch_size_per_image_so, positive_fraction_so) if bbox_reg_weights is None: bbox_reg_weights = (10., 10., 5., 5.) self.box_coder = det_utils.BoxCoder(bbox_reg_weights) self.box_roi_pool = box_roi_pool self.box_head = box_head self.box_predictor = box_predictor self.score_thresh = score_thresh self.nms_thresh = nms_thresh self.detections_per_img = detections_per_img self.mask_roi_pool = mask_roi_pool self.mask_head = mask_head self.mask_predictor = mask_predictor self.keypoint_roi_pool = keypoint_roi_pool self.keypoint_head = keypoint_head self.keypoint_predictor = keypoint_predictor
def __init__(self,
feature_extractor: Im2VecEncoder,
pooler_resolution: int = 7,
pooler_sampling_ratio: int = 2,
decoder_thresh: float = 0.1,
decoder_nms_thresh: float = 0.5,
decoder_detections_per_image: int = 100,
matcher_high_thresh: float = 0.5,
matcher_low_thresh: float = 0.5,
allow_low_quality_matches: bool = True,
batch_size_per_image: int = 256,
balance_sampling_fraction: float = 0.25):
super(FasterRCNNROIHead, self).__init__()
self.roi_pool = MultiScaleRoIAlign(
featmap_names=[0, 1, 2, 3],
output_size=pooler_resolution,
sampling_ratio=pooler_sampling_ratio)
self.feature_extractor = feature_extractor
self.box_coder = det_utils.BoxCoder(weights=(10., 10., 5., 5.))
self.decoder_thresh = decoder_thresh
self.decoder_nms_thresh = decoder_nms_thresh
self.decoder_detections_per_image = decoder_detections_per_image
self.proposal_matcher = det_utils.Matcher(
high_threshold=matcher_high_thresh,
low_threshold=matcher_low_thresh,
allow_low_quality_matches=allow_low_quality_matches)
self.sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction=balance_sampling_fraction)
def __init__(
self,
anchor_generator,
head,
#
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
#
pre_nms_top_n,
post_nms_top_n,
nms_thresh):
super(RegionProposalNetwork, self).__init__()
self.anchor_generator = anchor_generator
self.head = head
self.box_coder = det_utils.BoxCoder(weights=(1.0, 1.0, 1.0, 1.0))
# used during training
self.box_similarity = box_ops.box_iou
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh,
bg_iou_thresh,
allow_low_quality_matches=False,
)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
# used during testing
self._pre_nms_top_n = pre_nms_top_n
self._post_nms_top_n = post_nms_top_n
self.nms_thresh = nms_thresh
self.min_size = 1e-3
def __init__(self,
box_roi_pool,
box_head,
box_predictor,
# Faster R-CNN training
fg_iou_thresh, bg_iou_thresh,
batch_size_per_image, positive_fraction,
bbox_reg_weights,
# Faster R-CNN inference
score_thresh,
nms_thresh,
detections_per_img,
# Mask
mask_coarse_head=None,
mask_point_head=None,
):
super(PointRendRoIHeads, self).__init__()
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh,
bg_iou_thresh,
allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image,
positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_roi_pool = box_roi_pool
self.box_head = box_head
self.box_predictor = box_predictor
self.score_thresh = score_thresh
self.nms_thresh = nms_thresh
self.detections_per_img = detections_per_img
self.mask_coarse_in_features=("0",) #对应FPN特征图中的P2
self.mask_coarse_side_size=14
self._feature_scales={'0':0.25,'1':0.125,'2':0.0625,'3':0.03125,'4':0.015625} #FPN每阶段特征尺寸与原图像尺寸的比例
self.mask_coarse_head=mask_coarse_head
self.mask_point_in_features =["0"] #0对应FPN中的p2
self.mask_point_train_num_points =14*14
self.mask_point_oversample_ratio =3
self.mask_point_importance_sample_ratio =0.75
#next two parameters are use in the adaptive subdivions inference procedure
self.mask_point_subdivision_steps =5
self.mask_point_subdivision_num_points =28*28
self.mask_point_head=mask_point_head
def test_balanced_positive_negative_sampler(self):
sampler = _utils.BalancedPositiveNegativeSampler(4, 0.25)
# keep all 6 negatives first, then add 3 positives, last two are ignore
matched_idxs = [torch.tensor([0, 0, 0, 0, 0, 0, 1, 1, 1, -1, -1])]
pos, neg = sampler(matched_idxs)
# we know the number of elements that should be sampled for the positive (1)
# and the negative (3), and their location. Let's make sure that they are
# there
assert pos[0].sum() == 1
assert pos[0][6:9].sum() == 1
assert neg[0].sum() == 3
assert neg[0][0:6].sum() == 3
def __init__(self):
super(RoIHeads, self).__init__()
self.box_roi_pool = MultiScaleRoIAlign(
featmap_names=['0', '1', '2', '3'],
output_size=7,
sampling_ratio=2)
resolution = self.box_roi_pool.output_size[0]
representation_size = 1024
self.box_head = TwoMLPHead(256 * resolution**2, representation_size)
self.rlp_head = copy.deepcopy(self.box_head)
representation_size = 1024
self.box_predictor = FastRCNNPredictor(representation_size,
cfg.BOX.NUM_CLASSES)
self.RelDN = reldn_heads.reldn_head(self.box_head.fc7.out_features *
3) # concat of SPO
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
cfg.BOX.FG_IOU_THRESH,
cfg.BOX.BG_IOU_THRESH,
allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
cfg.BOX.BATCH_SIZE_PER_IMAGE, cfg.BOX.POSITIVE_FRACTION)
self.fg_bg_sampler_so = det_utils.BalancedPositiveNegativeSampler(
cfg.MODEL.BATCH_SIZE_PER_IMAGE_SO, cfg.MODEL.POSITIVE_FRACTION_SO)
self.fg_bg_sampler_rlp = det_utils.BalancedPositiveNegativeSampler(
cfg.MODEL.BATCH_SIZE_PER_IMAGE_REL,
cfg.MODEL.POSITIVE_FRACTION_REL)
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
def __init__(self,
backbone: ImageEncoder,
positive_fraction: float = 0.5,
match_thresh_low: float = 0.3,
match_thresh_high: float = 0.7,
anchor_sizes: List[int] = (128, 256, 512),
anchor_aspect_ratios: List[float] = (0.5, 1.0, 2.0),
batch_size_per_image: int = 256,
pre_nms_top_n: int = 6000,
post_nms_top_n: int = 300,
nms_thresh: float = 0.7,
min_size: int = 0,
fpn_post_nms_top_n: int = 1000,
fpn_post_nms_per_batch: int = True,
allow_low_quality_matches: bool = True,
initializer: InitializerApplicator = InitializerApplicator()) -> None:
super(RPN, self).__init__(None)
self._rpn_head = RPNHead(256, 3)
self.min_size = min_size
self.pre_nms_top_n = pre_nms_top_n
self.nms_thresh = nms_thresh
self.post_nms_top_n = post_nms_top_n
# the BoxCoder just converts the relative regression offsets into absolute
# coordinates
self.box_coder = det_utils.BoxCoder(weights=(1., 1., 1., 1.))
# sampler is responsible for selecting a subset of anchor boxes for computing the loss
# this makes sure each batch has reasonable balance of foreground/background labels
# it selects `batch_size_per_image` total boxes
self.sampler = det_utils.BalancedPositiveNegativeSampler(batch_size_per_image,
positive_fraction)
# matcher decides if an anchor box is a foreground or background based on how much
# it overlaps with the nearest target box
self.proposal_matcher = det_utils.Matcher(
match_thresh_high,
match_thresh_low,
allow_low_quality_matches=allow_low_quality_matches)
self.backbone = backbone
self.anchor_generator = AnchorGenerator(anchor_sizes, anchor_aspect_ratios)
self.num_anchors = self.anchor_generator.num_anchors_per_location()[0]
self._loss_meters = {'rpn_cls_loss': Average(), 'rpn_reg_loss': Average()}
initializer(self)
def __init__(
self,
box_roi_pool,
box_head,
box_predictor,
# Faster R-CNN training
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
bbox_reg_weights,
# new
weight_loss=False,
use_context=False,
track_embedding=None):
super(TrackHeads, self).__init__()
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_roi_pool = box_roi_pool
self.box_head = box_head
self.box_predictor = box_predictor
self.weight_loss = weight_loss
self.use_context = use_context
self.track_embedding = track_embedding
def __init__(
self,
box_predictor,
# Faster R-CNN training
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
bbox_reg_weights):
super(SSDHead, self).__init__()
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_predictor = box_predictor
def __init__(
self,
box_roi_pool,
box_head,
box_predictor,
# Faster R-CNN training
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
bbox_reg_weights,
# Faster R-CNN inference
score_thresh,
nms_thresh,
detections_per_img,
out_channels,
# Mask
mask_roi_pool=None,
mask_head=None,
mask_predictor=None,
keypoint_roi_pool=None,
keypoint_head=None,
keypoint_predictor=None,
pose_mean=None,
pose_stddev=None,
threed_68_points=None,
threed_5_points=None,
bbox_x_factor=1.1,
bbox_y_factor=1.1,
expand_forehead=0.3,
):
super(RoIHeads, self).__init__()
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10.0, 10.0, 5.0, 5.0)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_roi_pool = box_roi_pool
self.box_head = box_head
self.box_predictor = box_predictor
num_classes = 2
self.class_roi_pool = MultiScaleRoIAlign(
featmap_names=["0", "1", "2", "3"],
output_size=7,
sampling_ratio=2)
resolution = box_roi_pool.output_size[0]
representation_size = 1024
self.class_head = TwoMLPHead(out_channels * resolution**2,
representation_size)
self.class_predictor = FastRCNNClassPredictor(representation_size,
num_classes)
self.score_thresh = score_thresh
self.nms_thresh = nms_thresh
self.detections_per_img = detections_per_img
self.mask_roi_pool = mask_roi_pool
self.mask_head = mask_head
self.mask_predictor = mask_predictor
self.keypoint_roi_pool = keypoint_roi_pool
self.keypoint_head = keypoint_head
self.keypoint_predictor = keypoint_predictor
self.pose_mean = pose_mean
self.pose_stddev = pose_stddev
self.threed_68_points = threed_68_points
self.threed_5_points = threed_5_points
self.bbox_x_factor = bbox_x_factor
self.bbox_y_factor = bbox_y_factor
self.expand_forehead = expand_forehead
def __init__(
self,
box_roi_pool,
box_head,
box_predictor,
# Faster R-CNN training
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
bbox_reg_weights,
# Faster R-CNN inference
score_thresh,
nms_thresh,
detections_per_img,
# Mask
mask_roi_pool=None,
mask_head=None,
mask_predictor=None,
keypoint_roi_pool=None,
keypoint_head=None,
keypoint_predictor=None,
# new
use_soft_nms=False,
weight_loss=False,
use_context=False,
use_track_branch=False,
track_embedding=None):
super(RoIHeads, self).__init__()
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_roi_pool = box_roi_pool
self.box_head = box_head
self.box_predictor = box_predictor
self.score_thresh = score_thresh
self.nms_thresh = nms_thresh
self.detections_per_img = detections_per_img
self.mask_roi_pool = mask_roi_pool
self.mask_head = mask_head
self.mask_predictor = mask_predictor
self.keypoint_roi_pool = keypoint_roi_pool
self.keypoint_head = keypoint_head
self.keypoint_predictor = keypoint_predictor
self.weight_loss = weight_loss
self.use_soft_nms = use_soft_nms
self.use_context = use_context
def __init__(self,
out_channels,
num_classes,
input_mode,
acf_head,
fg_iou_thresh=0.5,
bg_iou_thresh=0.5,
batch_size_per_image=512,
positive_fraction=0.25,
bbox_reg_weights=None,
box_score_thresh=0.05,
box_nms_thresh=0.5,
box_detections_per_img=100):
super(RoIHeadsExtend, self).__init__()
self.in_channels = out_channels
self.input_mode = input_mode
self.score_thresh = box_score_thresh
self.nms_thresh = box_nms_thresh
self.detections_per_img = box_detections_per_img
self.fg_iou_thresh = fg_iou_thresh
self.bg_iou_thresh = bg_iou_thresh
self.batch_size_per_image = batch_size_per_image
self.positive_fraction = positive_fraction
self.num_classes = num_classes
# Detection
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh, bg_iou_thresh, allow_low_quality_matches=False)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = (10., 10., 5., 5.)
self.box_coder = det_utils.BoxCoder(bbox_reg_weights)
self.box_roi_pool = MultiScaleRoIAlign(featmap_names=[0, 1, 2, 3],
output_size=7,
sampling_ratio=2)
representation_size = 1024
resolution = self.box_roi_pool.output_size[0]
self.box_head = TwoMLPHead(out_channels * resolution**2,
representation_size)
self.box_predictor = FastRCNNPredictor(representation_size,
num_classes)
# Segmentation
self.shared_roi_pool = MultiScaleRoIAlign(featmap_names=[0, 1, 2, 3],
output_size=14,
sampling_ratio=2)
resolution = self.shared_roi_pool.output_size[0]
mask_layers = (256, 256, 256, 256, 256, 256, 256, 256)
mask_dilation = 1
self.mask_head = MaskRCNNHeads(out_channels, mask_layers,
mask_dilation)
mask_predictor_in_channels = 256 # == mask_layers[-1]
mask_dim_reduced = 256
self.mask_predictor = MaskRCNNPredictor(mask_predictor_in_channels,
mask_dim_reduced, num_classes)
self.with_paf_branch = True
if self.with_paf_branch:
self.paf_head = MaskRCNNHeads(out_channels, mask_layers,
mask_dilation)
self.paf_predictor = MaskRCNNPredictor(mask_predictor_in_channels,
mask_dim_reduced,
2 * (num_classes - 1))
if self.input_mode == config.INPUT_RGBD:
self.attention_block = ContextBlock(256, 2)
self.global_feature_dim = 256
self.with_3d_keypoints = True
self.with_axis_keypoints = False
self.regress_axis = False
self.estimate_norm_vector = False
if acf_head == 'endpoints':
self.with_axis_keypoints = True
elif acf_head == 'scatters':
self.regress_axis = True
elif acf_head == 'norm_vector':
self.estimate_norm_vector = True
else:
print("Don't assign a vaild acf head")
exit()
keypoint_layers = (256, ) * 4
self.keypoint_dim_reduced = keypoint_layers[-1]
if self.with_3d_keypoints:
self.vote_keypoint_head = Vote_Kpoints_head(
self.global_feature_dim, keypoint_layers, "conv2d")
self.vote_keypoint_predictor = Vote_Kpoints_Predictor(
self.keypoint_dim_reduced, 3 * (num_classes - 1))
if self.with_axis_keypoints:
self.orientation_keypoint_head = Vote_Kpoints_head(
self.global_feature_dim, keypoint_layers, "conv2d")
self.orientation_keypoint_predictor = Vote_Kpoints_Predictor(
self.keypoint_dim_reduced, 6 * (num_classes - 1))
if self.regress_axis:
self.axis_head = Vote_Kpoints_head(self.global_feature_dim,
keypoint_layers, "conv2d")
self.axis_predictor = Vote_Kpoints_Predictor(
self.keypoint_dim_reduced, 4 * (num_classes - 1))
if self.estimate_norm_vector:
self.norm_vector_head = Vote_Kpoints_head(
self.global_feature_dim, keypoint_layers, "conv2d")
self.norm_vector_predictor = Vote_Kpoints_Predictor(
self.keypoint_dim_reduced, 3 * (num_classes - 1))
def __init__(
self,
box_roi_pool,
box_head,
box_predictor,
# Faster R-CNN training
fg_iou_thresh,
bg_iou_thresh,
batch_size_per_image,
positive_fraction,
bbox_reg_weights,
# Faster R-CNN inference
score_thresh,
nms_thresh,
detections_per_img,
# Mask
mask_roi_pool=None,
mask_head=None,
mask_predictor=None,
keypoint_roi_pool=None,
keypoint_head=None,
keypoint_predictor=None,
):
super(CascadeRoIHeads, self).__init__()
self.num_cascade_stages = len(box_head)
self.box_similarity = box_ops.box_iou
# assign ground-truth boxes for each proposal
self.proposal_matcher = det_utils.Matcher(
fg_iou_thresh[0],
bg_iou_thresh[0],
allow_low_quality_matches=False)
self.proposal_matchers = []
for i in range(3):
proposal_matcher = det_utils.Matcher(
fg_iou_thresh[i],
bg_iou_thresh[i],
allow_low_quality_matches=False)
self.proposal_matchers.append(proposal_matcher)
self.fg_bg_sampler = det_utils.BalancedPositiveNegativeSampler(
batch_size_per_image, positive_fraction)
if bbox_reg_weights is None:
bbox_reg_weights = [(10., 10., 5., 5.), (20., 20., 10., 10.),
(30., 30., 15., 15.)]
self.box_coders = []
for i in range(3):
self.box_coders.append(det_utils.BoxCoder(bbox_reg_weights[i]))
self.box_roi_pool = box_roi_pool
self.box_head = box_head
self.box_predictor = box_predictor
self.score_thresh = score_thresh
self.nms_thresh = nms_thresh
self.detections_per_img = detections_per_img
self.mask_roi_pool = mask_roi_pool
self.mask_head = mask_head
self.mask_predictor = mask_predictor
self.keypoint_roi_pool = keypoint_roi_pool
self.keypoint_head = keypoint_head
self.keypoint_predictor = keypoint_predictor