Source code for mmrotate.models.dense_heads.rotated_reppoints_head
# Copyright (c) OpenMMLab. All rights reserved.
import numpy as np
import torch
import torch.nn as nn
from mmcv.cnn import ConvModule
from mmcv.ops import DeformConv2d, min_area_polygons
from mmcv.runner import force_fp32
from mmdet.core import images_to_levels, multi_apply, unmap
from mmdet.core.anchor.point_generator import MlvlPointGenerator
from mmdet.core.utils import select_single_mlvl
from mmdet.models.dense_heads.base_dense_head import BaseDenseHead
from mmrotate.core import (build_assigner, build_sampler,
multiclass_nms_rotated, obb2poly, poly2obb)
from ..builder import ROTATED_HEADS, build_loss
from .utils import convex_overlaps, levels_to_images
[docs]@ROTATED_HEADS.register_module()
class RotatedRepPointsHead(BaseDenseHead):
"""Rotated RepPoints head.
Args:
num_classes (int): Number of classes.
in_channels (int): Number of input channels.
feat_channels (int): Number of feature channels.
point_feat_channels (int, optional): Number of channels of points
features.
stacked_convs (int, optional): Number of stacked convolutions.
num_points (int, optional): Number of points in points set.
gradient_mul (float, optional): The multiplier to gradients from
points refinement and recognition.
point_strides (Iterable, optional): points strides.
point_base_scale (int, optional): Bbox scale for assigning labels.
conv_bias (str, optional): The bias of convolution.
loss_cls (dict, optional): Config of classification loss.
loss_bbox_init (dict, optional): Config of initial points loss.
loss_bbox_refine (dict, optional): Config of points loss in refinement.
conv_cfg (dict, optional): The config of convolution.
norm_cfg (dict, optional): The config of normlization.
train_cfg (dict, optional): The config of train.
test_cfg (dict, optional): The config of test.
center_init (bool, optional): Whether to use center point assignment.
transform_method (str, optional): The methods to transform RepPoints
to bbox.
use_reassign (bool, optional): Whether to reassign samples.
topk (int, optional): Number of the highest topk points. Defaults to 9.
anti_factor (float, optional): Feature anti-aliasing coefficient.
version (str, optional): Angle representations. Defaults to 'oc'.
init_cfg (dict or list[dict], optional): Initialization config dict.
"""
def __init__(self,
num_classes,
in_channels,
feat_channels,
point_feat_channels=256,
stacked_convs=3,
num_points=9,
gradient_mul=0.1,
point_strides=[8, 16, 32, 64, 128],
point_base_scale=4,
conv_bias='auto',
loss_cls=dict(
type='FocalLoss',
use_sigmoid=True,
gamma=2.0,
alpha=0.25,
loss_weight=1.0),
loss_bbox_init=dict(
type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=0.5),
loss_bbox_refine=dict(
type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0),
conv_cfg=None,
norm_cfg=None,
train_cfg=None,
test_cfg=None,
center_init=True,
transform_method='rotrect',
use_reassign=False,
topk=6,
anti_factor=0.75,
version='oc',
init_cfg=dict(
type='Normal',
layer='Conv2d',
std=0.01,
override=dict(
type='Normal',
name='reppoints_cls_out',
std=0.01,
bias_prob=0.01)),
**kwargs):
super(RotatedRepPointsHead, self).__init__(init_cfg)
self.num_points = num_points
self.point_feat_channels = point_feat_channels
self.center_init = center_init
# we use deform conv to extract points features
self.dcn_kernel = int(np.sqrt(num_points))
self.dcn_pad = int((self.dcn_kernel - 1) / 2)
assert self.dcn_kernel * self.dcn_kernel == num_points, \
'The points number should be a square number.'
assert self.dcn_kernel % 2 == 1, \
'The points number should be an odd square number.'
dcn_base = np.arange(-self.dcn_pad,
self.dcn_pad + 1).astype(np.float64)
dcn_base_y = np.repeat(dcn_base, self.dcn_kernel)
dcn_base_x = np.tile(dcn_base, self.dcn_kernel)
dcn_base_offset = np.stack([dcn_base_y, dcn_base_x], axis=1).reshape(
(-1))
self.dcn_base_offset = torch.tensor(dcn_base_offset).view(1, -1, 1, 1)
self.num_classes = num_classes
self.in_channels = in_channels
self.feat_channels = feat_channels
self.stacked_convs = stacked_convs
assert conv_bias == 'auto' or isinstance(conv_bias, bool)
self.conv_bias = conv_bias
self.loss_cls = build_loss(loss_cls)
self.train_cfg = train_cfg
self.test_cfg = test_cfg
self.conv_cfg = conv_cfg
self.norm_cfg = norm_cfg
self.fp16_enabled = False
self.gradient_mul = gradient_mul
self.point_base_scale = point_base_scale
self.point_strides = point_strides
self.prior_generator = MlvlPointGenerator(
self.point_strides, offset=0.)
self.num_base_priors = self.prior_generator.num_base_priors[0]
self.sampling = loss_cls['type'] not in ['FocalLoss']
if self.train_cfg:
self.init_assigner = build_assigner(self.train_cfg.init.assigner)
self.refine_assigner = build_assigner(
self.train_cfg.refine.assigner)
# use PseudoSampler when sampling is False
if self.sampling and hasattr(self.train_cfg, 'sampler'):
sampler_cfg = self.train_cfg.sampler
else:
sampler_cfg = dict(type='PseudoSampler')
self.sampler = build_sampler(sampler_cfg, context=self)
self.transform_method = transform_method
self.use_sigmoid_cls = loss_cls.get('use_sigmoid', False)
if self.use_sigmoid_cls:
self.cls_out_channels = self.num_classes
else:
self.cls_out_channels = self.num_classes + 1
self.loss_bbox_init = build_loss(loss_bbox_init)
self.loss_bbox_refine = build_loss(loss_bbox_refine)
self.use_reassign = use_reassign
self.topk = topk
self.anti_factor = anti_factor
self.version = version
self._init_layers()
def _init_layers(self):
"""Initialize layers of the head."""
self.relu = nn.ReLU(inplace=True)
self.cls_convs = nn.ModuleList()
self.reg_convs = nn.ModuleList()
for i in range(self.stacked_convs):
chn = self.in_channels if i == 0 else self.feat_channels
self.cls_convs.append(
ConvModule(
chn,
self.feat_channels,
3,
stride=1,
padding=1,
conv_cfg=self.conv_cfg,
norm_cfg=self.norm_cfg,
bias=self.conv_bias))
self.reg_convs.append(
ConvModule(
chn,
self.feat_channels,
3,
stride=1,
padding=1,
conv_cfg=self.conv_cfg,
norm_cfg=self.norm_cfg,
bias=self.conv_bias))
pts_out_dim = 2 * self.num_points
self.reppoints_cls_conv = DeformConv2d(self.feat_channels,
self.point_feat_channels,
self.dcn_kernel, 1,
self.dcn_pad)
self.reppoints_cls_out = nn.Conv2d(self.point_feat_channels,
self.cls_out_channels, 1, 1, 0)
self.reppoints_pts_init_conv = nn.Conv2d(self.feat_channels,
self.point_feat_channels, 3,
1, 1)
self.reppoints_pts_init_out = nn.Conv2d(self.point_feat_channels,
pts_out_dim, 1, 1, 0)
self.reppoints_pts_refine_conv = DeformConv2d(self.feat_channels,
self.point_feat_channels,
self.dcn_kernel, 1,
self.dcn_pad)
self.reppoints_pts_refine_out = nn.Conv2d(self.point_feat_channels,
pts_out_dim, 1, 1, 0)
[docs] def points2rotrect(self, pts, y_first=True):
"""Convert points to oriented bboxes."""
if y_first:
pts = pts.reshape(-1, self.num_points, 2)
pts_dy = pts[:, :, 0::2]
pts_dx = pts[:, :, 1::2]
pts = torch.cat([pts_dx, pts_dy],
dim=2).reshape(-1, 2 * self.num_points)
if self.transform_method == 'rotrect':
rotrect_pred = min_area_polygons(pts)
return rotrect_pred
else:
raise NotImplementedError
[docs] def forward(self, feats):
"""Forward function."""
return multi_apply(self.forward_single, feats)
[docs] def forward_single(self, x):
"""Forward feature map of a single FPN level."""
dcn_base_offset = self.dcn_base_offset.type_as(x)
points_init = 0
cls_feat = x
pts_feat = x
for cls_conv in self.cls_convs:
cls_feat = cls_conv(cls_feat)
for reg_conv in self.reg_convs:
pts_feat = reg_conv(pts_feat)
# initialize reppoints
pts_out_init = self.reppoints_pts_init_out(
self.relu(self.reppoints_pts_init_conv(pts_feat)))
pts_out_init = pts_out_init + points_init
# refine and classify reppoints
pts_out_init_grad_mul = (1 - self.gradient_mul) * pts_out_init.detach(
) + self.gradient_mul * pts_out_init
dcn_offset = pts_out_init_grad_mul - dcn_base_offset
cls_out = self.reppoints_cls_out(
self.relu(self.reppoints_cls_conv(cls_feat, dcn_offset)))
pts_out_refine = self.reppoints_pts_refine_out(
self.relu(self.reppoints_pts_refine_conv(pts_feat, dcn_offset)))
pts_out_refine = pts_out_refine + pts_out_init.detach()
return cls_out, pts_out_init, pts_out_refine
[docs] def get_points(self, featmap_sizes, img_metas, device):
"""Get points according to feature map sizes.
Args:
featmap_sizes (list[tuple]): Multi-level feature map sizes.
img_metas (list[dict]): Image meta info.
Returns:
tuple: points of each image, valid flags of each image
"""
num_imgs = len(img_metas)
multi_level_points = self.prior_generator.grid_priors(
featmap_sizes, device=device, with_stride=True)
points_list = [[point.clone() for point in multi_level_points]
for _ in range(num_imgs)]
valid_flag_list = []
for img_id, img_meta in enumerate(img_metas):
multi_level_flags = self.prior_generator.valid_flags(
featmap_sizes, img_meta['pad_shape'])
valid_flag_list.append(multi_level_flags)
return points_list, valid_flag_list
[docs] def offset_to_pts(self, center_list, pred_list):
"""Change from point offset to point coordinate."""
pts_list = []
for i_lvl, _ in enumerate(self.point_strides):
pts_lvl = []
for i_img, _ in enumerate(center_list):
pts_center = center_list[i_img][i_lvl][:, :2].repeat(
1, self.num_points)
pts_shift = pred_list[i_lvl][i_img]
yx_pts_shift = pts_shift.permute(1, 2, 0).view(
-1, 2 * self.num_points)
y_pts_shift = yx_pts_shift[..., 0::2]
x_pts_shift = yx_pts_shift[..., 1::2]
xy_pts_shift = torch.stack([x_pts_shift, y_pts_shift], -1)
xy_pts_shift = xy_pts_shift.view(*yx_pts_shift.shape[:-1], -1)
pts = xy_pts_shift * self.point_strides[i_lvl] + pts_center
pts_lvl.append(pts)
pts_lvl = torch.stack(pts_lvl, 0)
pts_list.append(pts_lvl)
return pts_list
def _point_target_single(self,
flat_proposals,
valid_flags,
gt_bboxes,
gt_bboxes_ignore,
gt_labels,
overlaps,
stage='init',
unmap_outputs=True):
"""Single point target function."""
inside_flags = valid_flags
if not inside_flags.any():
return (None, ) * 8
# assign gt and sample proposals
proposals = flat_proposals[inside_flags, :]
if stage == 'init':
assigner = self.init_assigner
pos_weight = self.train_cfg.init.pos_weight
else:
assigner = self.refine_assigner
pos_weight = self.train_cfg.refine.pos_weight
# convert gt from obb to poly
gt_bboxes = obb2poly(gt_bboxes, self.version)
assign_result = assigner.assign(proposals, gt_bboxes, overlaps,
gt_bboxes_ignore,
None if self.sampling else gt_labels)
sampling_result = self.sampler.sample(assign_result, proposals,
gt_bboxes)
num_valid_proposals = proposals.shape[0]
bbox_gt = proposals.new_zeros([num_valid_proposals, 8])
pos_proposals = torch.zeros_like(proposals)
proposals_weights = proposals.new_zeros(num_valid_proposals)
labels = proposals.new_full((num_valid_proposals, ),
self.num_classes,
dtype=torch.long)
label_weights = proposals.new_zeros(
num_valid_proposals, dtype=torch.float)
pos_inds = sampling_result.pos_inds
neg_inds = sampling_result.neg_inds
if len(pos_inds) > 0:
pos_gt_bboxes = sampling_result.pos_gt_bboxes
bbox_gt[pos_inds, :] = pos_gt_bboxes
pos_proposals[pos_inds, :] = proposals[pos_inds, :]
proposals_weights[pos_inds] = 1.0
if gt_labels is None:
# Only rpn gives gt_labels as None
# Foreground is the first class
labels[pos_inds] = 0
else:
labels[pos_inds] = gt_labels[
sampling_result.pos_assigned_gt_inds]
if pos_weight <= 0:
label_weights[pos_inds] = 1.0
else:
label_weights[pos_inds] = pos_weight
if len(neg_inds) > 0:
label_weights[neg_inds] = 1.0
# map up to original set of proposals
if unmap_outputs:
num_total_proposals = flat_proposals.size(0)
labels = unmap(labels, num_total_proposals, inside_flags)
label_weights = unmap(label_weights, num_total_proposals,
inside_flags)
bbox_gt = unmap(bbox_gt, num_total_proposals, inside_flags)
pos_proposals = unmap(pos_proposals, num_total_proposals,
inside_flags)
proposals_weights = unmap(proposals_weights, num_total_proposals,
inside_flags)
return (labels, label_weights, bbox_gt, pos_proposals,
proposals_weights, pos_inds, neg_inds, sampling_result)
[docs] def get_targets(self,
proposals_list,
valid_flag_list,
gt_bboxes_list,
img_metas,
gt_bboxes_ignore_list=None,
gt_labels_list=None,
stage='init',
label_channels=1,
unmap_outputs=True):
"""Compute corresponding GT box and classification targets for
proposals.
Args:
proposals_list (list[list]): Multi level points/bboxes of each
image.
valid_flag_list (list[list]): Multi level valid flags of each
image.
gt_bboxes_list (list[Tensor]): Ground truth bboxes of each image.
img_metas (list[dict]): Meta info of each image.
gt_bboxes_ignore_list (list[Tensor]): Ground truth bboxes to be
ignored.
gt_bboxes_list (list[Tensor]): Ground truth labels of each box.
stage (str): `init` or `refine`. Generate target for init stage or
refine stage
label_channels (int): Channel of label.
unmap_outputs (bool): Whether to map outputs back to the original
set of anchors.
Returns:
tuple (list[Tensor]):
- labels_list (list[Tensor]): Labels of each level.
- label_weights_list (list[Tensor]): Label weights of each \
level.
- bbox_gt_list (list[Tensor]): Ground truth bbox of each level.
- proposal_list (list[Tensor]): Proposals(points/bboxes) of \
each level.
- proposal_weights_list (list[Tensor]): Proposal weights of \
each level.
- num_total_pos (int): Number of positive samples in all \
images.
- num_total_neg (int): Number of negative samples in all \
images.
"""
assert stage in ['init', 'refine']
num_imgs = len(img_metas)
assert len(proposals_list) == len(valid_flag_list) == num_imgs
# points number of multi levels
num_level_proposals = [points.size(0) for points in proposals_list[0]]
# concat all level points and flags to a single tensor
for i in range(num_imgs):
assert len(proposals_list[i]) == len(valid_flag_list[i])
proposals_list[i] = torch.cat(proposals_list[i])
valid_flag_list[i] = torch.cat(valid_flag_list[i])
# compute targets for each image
if gt_bboxes_ignore_list is None:
gt_bboxes_ignore_list = [None for _ in range(num_imgs)]
if gt_labels_list is None:
gt_labels_list = [None for _ in range(num_imgs)]
all_overlaps_rotate_list = [None] * 4
(all_labels, all_label_weights, all_bbox_gt, all_proposals,
all_proposal_weights, pos_inds_list, neg_inds_list,
sampling_result) = multi_apply(
self._point_target_single,
proposals_list,
valid_flag_list,
gt_bboxes_list,
gt_bboxes_ignore_list,
gt_labels_list,
all_overlaps_rotate_list,
stage=stage,
unmap_outputs=unmap_outputs)
# no valid points
if any([labels is None for labels in all_labels]):
return None
# sampled points of all images
num_total_pos = sum([max(inds.numel(), 1) for inds in pos_inds_list])
num_total_neg = sum([max(inds.numel(), 1) for inds in neg_inds_list])
labels_list = images_to_levels(all_labels, num_level_proposals)
label_weights_list = images_to_levels(all_label_weights,
num_level_proposals)
bbox_gt_list = images_to_levels(all_bbox_gt, num_level_proposals)
proposals_list = images_to_levels(all_proposals, num_level_proposals)
proposal_weights_list = images_to_levels(all_proposal_weights,
num_level_proposals)
return (labels_list, label_weights_list, bbox_gt_list, proposals_list,
proposal_weights_list, num_total_pos, num_total_neg, None)
[docs] def get_cfa_targets(self,
proposals_list,
valid_flag_list,
gt_bboxes_list,
img_metas,
gt_bboxes_ignore_list=None,
gt_labels_list=None,
stage='init',
label_channels=1,
unmap_outputs=True):
"""Compute corresponding GT box and classification targets for
proposals.
Args:
proposals_list (list[list]): Multi level points/bboxes of each
image.
valid_flag_list (list[list]): Multi level valid flags of each
image.
gt_bboxes_list (list[Tensor]): Ground truth bboxes of each image.
img_metas (list[dict]): Meta info of each image.
gt_bboxes_ignore_list (list[Tensor]): Ground truth bboxes to be
ignored.
gt_bboxes_list (list[Tensor]): Ground truth labels of each box.
stage (str): `init` or `refine`. Generate target for init stage or
refine stage
label_channels (int): Channel of label.
unmap_outputs (bool): Whether to map outputs back to the original
set of anchors.
Returns:
tuple:
- all_labels (list[Tensor]): Labels of each level.
- all_label_weights (list[Tensor]): Label weights of each \
level.
- all_bbox_gt (list[Tensor]): Ground truth bbox of each level.
- all_proposals (list[Tensor]): Proposals(points/bboxes) of \
each level.
- all_proposal_weights (list[Tensor]): Proposal weights of \
each level.
- pos_inds (list[Tensor]): Index of positive samples in all \
images.
- gt_inds (list[Tensor]): Index of ground truth bbox in all \
images.
"""
assert stage in ['init', 'refine']
num_imgs = len(img_metas)
assert len(proposals_list) == len(valid_flag_list) == num_imgs
# concat all level points and flags to a single tensor
for i in range(num_imgs):
assert len(proposals_list[i]) == len(valid_flag_list[i])
proposals_list[i] = torch.cat(proposals_list[i])
valid_flag_list[i] = torch.cat(valid_flag_list[i])
# compute targets for each image
if gt_bboxes_ignore_list is None:
gt_bboxes_ignore_list = [None for _ in range(num_imgs)]
if gt_labels_list is None:
gt_labels_list = [None for _ in range(num_imgs)]
all_overlaps_rotate_list = [None] * 4
(all_labels, all_label_weights, all_bbox_gt, all_proposals,
all_proposal_weights, pos_inds_list, neg_inds_list,
sampling_result) = multi_apply(
self._point_target_single,
proposals_list,
valid_flag_list,
gt_bboxes_list,
gt_bboxes_ignore_list,
gt_labels_list,
all_overlaps_rotate_list,
stage=stage,
unmap_outputs=unmap_outputs)
pos_inds = []
# pos_gt_index = []
for i, single_labels in enumerate(all_labels):
pos_mask = (0 <= single_labels) & (
single_labels < self.num_classes)
pos_inds.append(pos_mask.nonzero(as_tuple=False).view(-1))
gt_inds = [item.pos_assigned_gt_inds for item in sampling_result]
return (all_labels, all_label_weights, all_bbox_gt, all_proposals,
all_proposal_weights, pos_inds, gt_inds)
[docs] def loss_single(self, cls_score, pts_pred_init, pts_pred_refine, labels,
label_weights, rbbox_gt_init, convex_weights_init,
rbbox_gt_refine, convex_weights_refine, stride,
num_total_samples_refine):
"""Single loss function."""
normalize_term = self.point_base_scale * stride
if self.use_reassign:
rbbox_gt_init = rbbox_gt_init.reshape(-1, 8)
convex_weights_init = convex_weights_init.reshape(-1)
pts_pred_init = pts_pred_init.reshape(-1, 2 * self.num_points)
pos_ind_init = (convex_weights_init > 0).nonzero(
as_tuple=False).reshape(-1)
pts_pred_init_norm = pts_pred_init[pos_ind_init]
rbbox_gt_init_norm = rbbox_gt_init[pos_ind_init]
convex_weights_pos_init = convex_weights_init[pos_ind_init]
loss_pts_init = self.loss_bbox_init(
pts_pred_init_norm / normalize_term,
rbbox_gt_init_norm / normalize_term, convex_weights_pos_init)
return 0, loss_pts_init, 0
else:
rbbox_gt_init = rbbox_gt_init.reshape(-1, 8)
convex_weights_init = convex_weights_init.reshape(-1)
# init points loss
pts_pred_init = pts_pred_init.reshape(-1, 2 * self.num_points)
pos_ind_init = (convex_weights_init > 0).nonzero(
as_tuple=False).reshape(-1)
pts_pred_init_norm = pts_pred_init[pos_ind_init]
rbbox_gt_init_norm = rbbox_gt_init[pos_ind_init]
convex_weights_pos_init = convex_weights_init[pos_ind_init]
loss_pts_init = self.loss_bbox_init(
pts_pred_init_norm / normalize_term,
rbbox_gt_init_norm / normalize_term, convex_weights_pos_init)
# refine points loss
rbbox_gt_refine = rbbox_gt_refine.reshape(-1, 8)
pts_pred_refine = pts_pred_refine.reshape(-1, 2 * self.num_points)
convex_weights_refine = convex_weights_refine.reshape(-1)
pos_ind_refine = (convex_weights_refine > 0).nonzero(
as_tuple=False).reshape(-1)
pts_pred_refine_norm = pts_pred_refine[pos_ind_refine]
rbbox_gt_refine_norm = rbbox_gt_refine[pos_ind_refine]
convex_weights_pos_refine = convex_weights_refine[pos_ind_refine]
loss_pts_refine = self.loss_bbox_refine(
pts_pred_refine_norm / normalize_term,
rbbox_gt_refine_norm / normalize_term,
convex_weights_pos_refine)
# classification loss
labels = labels.reshape(-1)
label_weights = label_weights.reshape(-1)
cls_score = cls_score.permute(0, 2, 3,
1).reshape(-1, self.cls_out_channels)
loss_cls = self.loss_cls(
cls_score,
labels,
label_weights,
avg_factor=num_total_samples_refine)
return loss_cls, loss_pts_init, loss_pts_refine
[docs] def loss(self,
cls_scores,
pts_preds_init,
pts_preds_refine,
gt_bboxes,
gt_labels,
img_metas,
gt_bboxes_ignore=None):
"""Loss function of CFA head."""
featmap_sizes = [featmap.size()[-2:] for featmap in cls_scores]
assert len(featmap_sizes) == self.prior_generator.num_levels
label_channels = self.cls_out_channels if self.use_sigmoid_cls else 1
device = cls_scores[0].device
# target for initial stage
center_list, valid_flag_list = self.get_points(
featmap_sizes, img_metas, device=device)
pts_coordinate_preds_init = self.offset_to_pts(center_list,
pts_preds_init)
if self.use_reassign: # get num_proposal_each_lvl and lvl_num
num_proposals_each_level = [(featmap.size(-1) * featmap.size(-2))
for featmap in cls_scores]
num_level = len(featmap_sizes)
assert num_level == len(pts_coordinate_preds_init)
if self.train_cfg.init.assigner['type'] == 'ConvexAssigner':
candidate_list = center_list
else:
raise NotImplementedError
cls_reg_targets_init = self.get_targets(
candidate_list,
valid_flag_list,
gt_bboxes,
img_metas,
gt_bboxes_ignore_list=gt_bboxes_ignore,
gt_labels_list=gt_labels,
stage='init',
label_channels=label_channels)
(*_, rbbox_gt_list_init, candidate_list_init, convex_weights_list_init,
num_total_pos_init, num_total_neg_init, _) = cls_reg_targets_init
# target for refinement stage
center_list, valid_flag_list = self.get_points(
featmap_sizes, img_metas, device=device)
pts_coordinate_preds_refine = self.offset_to_pts(
center_list, pts_preds_refine)
points_list = []
for i_img, center in enumerate(center_list):
points = []
for i_lvl in range(len(pts_preds_refine)):
points_preds_init_ = pts_preds_init[i_lvl].detach()
points_preds_init_ = points_preds_init_.view(
points_preds_init_.shape[0], -1,
*points_preds_init_.shape[2:])
points_shift = points_preds_init_.permute(
0, 2, 3, 1) * self.point_strides[i_lvl]
points_center = center[i_lvl][:, :2].repeat(1, self.num_points)
points.append(
points_center +
points_shift[i_img].reshape(-1, 2 * self.num_points))
points_list.append(points)
if self.use_reassign:
cls_reg_targets_refine = self.get_cfa_targets(
points_list,
valid_flag_list,
gt_bboxes,
img_metas,
gt_bboxes_ignore_list=gt_bboxes_ignore,
gt_labels_list=gt_labels,
stage='refine',
label_channels=label_channels)
(labels_list, label_weights_list, rbbox_gt_list_refine, _,
convex_weights_list_refine, pos_inds_list_refine,
pos_gt_index_list_refine) = cls_reg_targets_refine
cls_scores = levels_to_images(cls_scores)
cls_scores = [
item.reshape(-1, self.cls_out_channels) for item in cls_scores
]
pts_coordinate_preds_init_cfa = levels_to_images(
pts_coordinate_preds_init, flatten=True)
pts_coordinate_preds_init_cfa = [
item.reshape(-1, 2 * self.num_points)
for item in pts_coordinate_preds_init_cfa
]
pts_coordinate_preds_refine = levels_to_images(
pts_coordinate_preds_refine, flatten=True)
pts_coordinate_preds_refine = [
item.reshape(-1, 2 * self.num_points)
for item in pts_coordinate_preds_refine
]
with torch.no_grad():
pos_losses_list, = multi_apply(
self.get_pos_loss, cls_scores,
pts_coordinate_preds_init_cfa, labels_list,
rbbox_gt_list_refine, label_weights_list,
convex_weights_list_refine, pos_inds_list_refine)
labels_list, label_weights_list, convex_weights_list_refine, \
num_pos, pos_normalize_term = multi_apply(
self.reassign,
pos_losses_list,
labels_list,
label_weights_list,
pts_coordinate_preds_init_cfa,
convex_weights_list_refine,
gt_bboxes,
pos_inds_list_refine,
pos_gt_index_list_refine,
num_proposals_each_level=num_proposals_each_level,
num_level=num_level
)
num_pos = sum(num_pos)
# convert all tensor list to a flatten tensor
cls_scores = torch.cat(cls_scores, 0).view(-1,
cls_scores[0].size(-1))
pts_preds_refine = torch.cat(pts_coordinate_preds_refine, 0).view(
-1, pts_coordinate_preds_refine[0].size(-1))
labels = torch.cat(labels_list, 0).view(-1)
labels_weight = torch.cat(label_weights_list, 0).view(-1)
rbbox_gt_refine = torch.cat(rbbox_gt_list_refine, 0).view(
-1, rbbox_gt_list_refine[0].size(-1))
convex_weights_refine = torch.cat(convex_weights_list_refine,
0).view(-1)
pos_normalize_term = torch.cat(pos_normalize_term, 0).reshape(-1)
pos_inds_flatten = ((0 <= labels) &
(labels < self.num_classes)).nonzero(
as_tuple=False).reshape(-1)
assert len(pos_normalize_term) == len(pos_inds_flatten)
if num_pos:
losses_cls = self.loss_cls(
cls_scores, labels, labels_weight, avg_factor=num_pos)
pos_pts_pred_refine = pts_preds_refine[pos_inds_flatten]
pos_rbbox_gt_refine = rbbox_gt_refine[pos_inds_flatten]
pos_convex_weights_refine = convex_weights_refine[
pos_inds_flatten]
losses_pts_refine = self.loss_bbox_refine(
pos_pts_pred_refine / pos_normalize_term.reshape(-1, 1),
pos_rbbox_gt_refine / pos_normalize_term.reshape(-1, 1),
pos_convex_weights_refine)
else:
losses_cls = cls_scores.sum() * 0
losses_pts_refine = pts_preds_refine.sum() * 0
None_list = [None] * num_level
_, losses_pts_init, _ = multi_apply(
self.loss_single,
None_list,
pts_coordinate_preds_init,
None_list,
None_list,
None_list,
rbbox_gt_list_init,
convex_weights_list_init,
None_list,
None_list,
self.point_strides,
num_total_samples_refine=None,
)
loss_dict_all = {
'loss_cls': losses_cls,
'loss_pts_init': losses_pts_init,
'loss_pts_refine': losses_pts_refine
}
return loss_dict_all
else:
cls_reg_targets_refine = self.get_targets(
points_list,
valid_flag_list,
gt_bboxes,
img_metas,
gt_bboxes_ignore_list=gt_bboxes_ignore,
gt_labels_list=gt_labels,
stage='refine',
label_channels=label_channels)
(labels_list, label_weights_list, rbbox_gt_list_refine,
candidate_list_refine, convex_weights_list_refine,
num_total_pos_refine, num_total_neg_refine,
_) = cls_reg_targets_refine
num_total_samples_refine = (
num_total_pos_refine + num_total_neg_refine
if self.sampling else num_total_pos_refine)
losses_cls, losses_pts_init, losses_pts_refine = multi_apply(
self.loss_single,
cls_scores,
pts_coordinate_preds_init,
pts_coordinate_preds_refine,
labels_list,
label_weights_list,
rbbox_gt_list_init,
convex_weights_list_init,
rbbox_gt_list_refine,
convex_weights_list_refine,
self.point_strides,
num_total_samples_refine=num_total_samples_refine)
loss_dict_all = {
'loss_cls': losses_cls,
'loss_pts_init': losses_pts_init,
'loss_pts_refine': losses_pts_refine
}
return loss_dict_all
[docs] def get_pos_loss(self, cls_score, pts_pred, label, bbox_gt, label_weight,
convex_weight, pos_inds):
"""Calculate loss of all potential positive samples obtained from first
match process.
Args:
cls_score (Tensor): Box scores of single image with shape
(num_anchors, num_classes)
pts_pred (Tensor): Box energies / deltas of single image
with shape (num_anchors, 4)
label (Tensor): classification target of each anchor with
shape (num_anchors,)
bbox_gt (Tensor): Ground truth box.
label_weight (Tensor): Classification loss weight of each
anchor with shape (num_anchors).
convex_weight (Tensor): Bbox weight of each anchor with shape
(num_anchors, 4).
pos_inds (Tensor): Index of all positive samples got from
first assign process.
Returns:
Tensor: Losses of all positive samples in single image.
"""
pos_scores = cls_score[pos_inds]
pos_pts_pred = pts_pred[pos_inds]
pos_bbox_gt = bbox_gt[pos_inds]
pos_label = label[pos_inds]
pos_label_weight = label_weight[pos_inds]
pos_convex_weight = convex_weight[pos_inds]
loss_cls = self.loss_cls(
pos_scores,
pos_label,
pos_label_weight,
avg_factor=self.loss_cls.loss_weight,
reduction_override='none')
loss_bbox = self.loss_bbox_refine(
pos_pts_pred,
pos_bbox_gt,
pos_convex_weight,
avg_factor=self.loss_cls.loss_weight,
reduction_override='none')
loss_cls = loss_cls.sum(-1)
pos_loss = loss_bbox + loss_cls
return pos_loss,
[docs] def reassign(self,
pos_losses,
label,
label_weight,
pts_pred_init,
convex_weight,
gt_bbox,
pos_inds,
pos_gt_inds,
num_proposals_each_level=None,
num_level=None):
"""CFA reassign process.
Args:
pos_losses (Tensor): Losses of all positive samples in
single image.
label (Tensor): classification target of each anchor with
shape (num_anchors,)
label_weight (Tensor): Classification loss weight of each
anchor with shape (num_anchors).
pts_pred_init (Tensor):
convex_weight (Tensor): Bbox weight of each anchor with shape
(num_anchors, 4).
gt_bbox (Tensor): Ground truth box.
pos_inds (Tensor): Index of all positive samples got from
first assign process.
pos_gt_inds (Tensor): Gt_index of all positive samples got
from first assign process.
num_proposals_each_level (list, optional): Number of proposals
of each level.
num_level (int, optional): Number of level.
Returns:
tuple: Usually returns a tuple containing learning targets.
- label (Tensor): classification target of each anchor after \
paa assign, with shape (num_anchors,)
- label_weight (Tensor): Classification loss weight of each \
anchor after paa assign, with shape (num_anchors).
- convex_weight (Tensor): Bbox weight of each anchor with \
shape (num_anchors, 4).
- pos_normalize_term (list): pos normalize term for refine \
points losses.
"""
if len(pos_inds) == 0:
return label, label_weight, convex_weight, 0, torch.tensor(
[]).type_as(convex_weight)
num_gt = pos_gt_inds.max() + 1
num_proposals_each_level_ = num_proposals_each_level.copy()
num_proposals_each_level_.insert(0, 0)
inds_level_interval = np.cumsum(num_proposals_each_level_)
pos_level_mask = []
for i in range(num_level):
mask = (pos_inds >= inds_level_interval[i]) & (
pos_inds < inds_level_interval[i + 1])
pos_level_mask.append(mask)
# convert gt from obb to poly
gt_bbox = obb2poly(gt_bbox, self.version)
overlaps_matrix = convex_overlaps(gt_bbox, pts_pred_init)
pos_inds_after_cfa = []
ignore_inds_after_cfa = []
re_assign_weights_after_cfa = []
for gt_ind in range(num_gt):
pos_inds_cfa = []
pos_loss_cfa = []
pos_overlaps_init_cfa = []
gt_mask = pos_gt_inds == gt_ind
for level in range(num_level):
level_mask = pos_level_mask[level]
level_gt_mask = level_mask & gt_mask
value, topk_inds = pos_losses[level_gt_mask].topk(
min(level_gt_mask.sum(), self.topk), largest=False)
pos_inds_cfa.append(pos_inds[level_gt_mask][topk_inds])
pos_loss_cfa.append(value)
pos_overlaps_init_cfa.append(
overlaps_matrix[:, pos_inds[level_gt_mask][topk_inds]])
pos_inds_cfa = torch.cat(pos_inds_cfa)
pos_loss_cfa = torch.cat(pos_loss_cfa)
pos_overlaps_init_cfa = torch.cat(pos_overlaps_init_cfa, 1)
if len(pos_inds_cfa) < 2:
pos_inds_after_cfa.append(pos_inds_cfa)
ignore_inds_after_cfa.append(pos_inds_cfa.new_tensor([]))
re_assign_weights_after_cfa.append(
pos_loss_cfa.new_ones([len(pos_inds_cfa)]))
else:
pos_loss_cfa, sort_inds = pos_loss_cfa.sort()
pos_inds_cfa = pos_inds_cfa[sort_inds]
pos_overlaps_init_cfa = pos_overlaps_init_cfa[:, sort_inds] \
.reshape(-1, len(pos_inds_cfa))
pos_loss_cfa = pos_loss_cfa.reshape(-1)
loss_mean = pos_loss_cfa.mean()
loss_var = pos_loss_cfa.var()
gauss_prob_density = \
(-(pos_loss_cfa - loss_mean) ** 2 / loss_var) \
.exp() / loss_var.sqrt()
index_inverted, _ = torch.arange(
len(gauss_prob_density)).sort(descending=True)
gauss_prob_inverted = torch.cumsum(
gauss_prob_density[index_inverted], 0)
gauss_prob = gauss_prob_inverted[index_inverted]
gauss_prob_norm = (gauss_prob - gauss_prob.min()) / \
(gauss_prob.max() - gauss_prob.min())
# splitting by gradient consistency
loss_curve = gauss_prob_norm * pos_loss_cfa
_, max_thr = loss_curve.topk(1)
reweights = gauss_prob_norm[:max_thr + 1]
# feature anti-aliasing coefficient
pos_overlaps_init_cfa = pos_overlaps_init_cfa[:, :max_thr + 1]
overlaps_level = pos_overlaps_init_cfa[gt_ind] / (
pos_overlaps_init_cfa.sum(0) + 1e-6)
reweights = \
self.anti_factor * overlaps_level * reweights + \
1e-6
re_assign_weights = \
reweights.reshape(-1) / reweights.sum() * \
torch.ones(len(reweights)).type_as(
gauss_prob_norm).sum()
pos_inds_temp = pos_inds_cfa[:max_thr + 1]
ignore_inds_temp = pos_inds_cfa.new_tensor([])
pos_inds_after_cfa.append(pos_inds_temp)
ignore_inds_after_cfa.append(ignore_inds_temp)
re_assign_weights_after_cfa.append(re_assign_weights)
pos_inds_after_cfa = torch.cat(pos_inds_after_cfa)
ignore_inds_after_cfa = torch.cat(ignore_inds_after_cfa)
re_assign_weights_after_cfa = torch.cat(re_assign_weights_after_cfa)
reassign_mask = (pos_inds.unsqueeze(1) != pos_inds_after_cfa).all(1)
reassign_ids = pos_inds[reassign_mask]
label[reassign_ids] = self.num_classes
label_weight[ignore_inds_after_cfa] = 0
convex_weight[reassign_ids] = 0
num_pos = len(pos_inds_after_cfa)
re_assign_weights_mask = (
pos_inds.unsqueeze(1) == pos_inds_after_cfa).any(1)
reweight_ids = pos_inds[re_assign_weights_mask]
label_weight[reweight_ids] = re_assign_weights_after_cfa
convex_weight[reweight_ids] = re_assign_weights_after_cfa
pos_level_mask_after_cfa = []
for i in range(num_level):
mask = (pos_inds_after_cfa >= inds_level_interval[i]) & (
pos_inds_after_cfa < inds_level_interval[i + 1])
pos_level_mask_after_cfa.append(mask)
pos_level_mask_after_cfa = torch.stack(pos_level_mask_after_cfa,
0).type_as(label)
pos_normalize_term = pos_level_mask_after_cfa * (
self.point_base_scale *
torch.as_tensor(self.point_strides).type_as(label)).reshape(-1, 1)
pos_normalize_term = pos_normalize_term[
pos_normalize_term > 0].type_as(convex_weight)
assert len(pos_normalize_term) == len(pos_inds_after_cfa)
return label, label_weight, convex_weight, num_pos, pos_normalize_term
[docs] @force_fp32(apply_to=('cls_scores', 'pts_preds_init', 'pts_preds_refine'))
def get_bboxes(self,
cls_scores,
pts_preds_init,
pts_preds_refine,
img_metas,
cfg=None,
rescale=False,
with_nms=True,
**kwargs):
"""Transform network outputs of a batch into bbox results.
Args:
cls_scores (list[Tensor]): Classification scores for all
scale levels, each is a 4D-tensor, has shape
(batch_size, num_priors * num_classes, H, W).
pts_preds_init (list[Tensor]): Box energies / deltas for all
scale levels, each is a 18D-tensor, has shape
(batch_size, num_points * 2, H, W).
pts_preds_refine (list[Tensor]): Box energies / deltas for all
scale levels, each is a 18D-tensor, has shape
(batch_size, num_points * 2, H, W).
img_metas (list[dict], Optional): Image meta info. Default None.
cfg (mmcv.Config, Optional): Test / postprocessing configuration,
if None, test_cfg would be used. Default None.
rescale (bool): If True, return boxes in original image space.
Default False.
with_nms (bool): If True, do nms before return boxes.
Default True.
Returns:
list[list[Tensor, Tensor]]: Each item in result_list is 2-tuple.
The first item is an (n, 6) tensor, where the first 4 columns
are bounding box positions (cx, cy, w, h, a) and the
6-th column is a score between 0 and 1. The second item is a
(n,) tensor where each item is the predicted class label of
the corresponding box.
"""
assert len(cls_scores) == len(pts_preds_refine)
num_levels = len(cls_scores)
featmap_sizes = [cls_scores[i].shape[-2:] for i in range(num_levels)]
mlvl_priors = self.prior_generator.grid_priors(
featmap_sizes,
dtype=cls_scores[0].device,
device=cls_scores[0].device)
result_list = []
for img_id, _ in enumerate(img_metas):
img_meta = img_metas[img_id]
cls_score_list = select_single_mlvl(cls_scores, img_id)
point_pred_list = select_single_mlvl(pts_preds_refine, img_id)
results = self._get_bboxes_single(cls_score_list, point_pred_list,
mlvl_priors, img_meta, cfg,
rescale, with_nms, **kwargs)
result_list.append(results)
return result_list
def _get_bboxes_single(self,
cls_score_list,
point_pred_list,
mlvl_priors,
img_meta,
cfg,
rescale=False,
with_nms=True,
**kwargs):
"""Transform outputs of a single image into bbox predictions.
Args:
cls_score_list (list[Tensor]): Box scores from all scale
levels of a single image, each item has shape
(num_priors * num_classes, H, W).
bbox_pred_list (list[Tensor]): Box energies / deltas from
all scale levels of a single image, each item has shape
(num_priors * 4, H, W).
score_factor_list (list[Tensor]): Score factor from all scale
levels of a single image. RepPoints head does not need
this value.
mlvl_priors (list[Tensor]): Each element in the list is
the priors of a single level in feature pyramid, has shape
(num_priors, 2).
img_meta (dict): Image meta info.
cfg (mmcv.Config): Test / postprocessing configuration,
if None, test_cfg would be used.
rescale (bool): If True, return boxes in original image space.
Default: False.
with_nms (bool): If True, do nms before return boxes.
Default: True.
Returns:
tuple[Tensor]: Results of detected bboxes and labels. If with_nms
is False and mlvl_score_factor is None, return mlvl_bboxes and
mlvl_scores, else return mlvl_bboxes, mlvl_scores and
mlvl_score_factor. Usually with_nms is False is used for aug
test. If with_nms is True, then return the following format
- det_bboxes (Tensor): Predicted bboxes with shape \
[num_bboxes, 5], where the first 4 columns are bounding \
box positions (cx, cy, w, h, a) and the 5-th \
column are scores between 0 and 1.
- det_labels (Tensor): Predicted labels of the corresponding \
box with shape [num_bboxes].
"""
cfg = self.test_cfg if cfg is None else cfg
assert len(cls_score_list) == len(point_pred_list)
scale_factor = img_meta['scale_factor']
mlvl_bboxes = []
mlvl_scores = []
for level_idx, (cls_score, points_pred, points) in enumerate(
zip(cls_score_list, point_pred_list, mlvl_priors)):
assert cls_score.size()[-2:] == points_pred.size()[-2:]
cls_score = cls_score.permute(1, 2,
0).reshape(-1, self.cls_out_channels)
if self.use_sigmoid_cls:
scores = cls_score.sigmoid()
else:
scores = cls_score.softmax(-1)[:, :-1]
points_pred = points_pred.permute(1, 2, 0).reshape(
-1, 2 * self.num_points)
nms_pre = cfg.get('nms_pre', -1)
if 0 < nms_pre < scores.shape[0]:
if self.use_sigmoid_cls:
max_scores, _ = scores.max(dim=1)
else:
max_scores, _ = scores[:, 1:].max(dim=1)
_, topk_inds = max_scores.topk(nms_pre)
points = points[topk_inds, :]
points_pred = points_pred[topk_inds, :]
scores = scores[topk_inds, :]
poly_pred = self.points2rotrect(points_pred, y_first=True)
bbox_pos_center = points[:, :2].repeat(1, 4)
polys = poly_pred * self.point_strides[level_idx] + bbox_pos_center
bboxes = poly2obb(polys, self.version)
mlvl_bboxes.append(bboxes)
mlvl_scores.append(scores)
mlvl_bboxes = torch.cat(mlvl_bboxes)
if rescale:
mlvl_bboxes[..., :4] /= mlvl_bboxes[..., :4].new_tensor(
scale_factor)
mlvl_scores = torch.cat(mlvl_scores)
if self.use_sigmoid_cls:
padding = mlvl_scores.new_zeros(mlvl_scores.shape[0], 1)
mlvl_scores = torch.cat([mlvl_scores, padding], dim=1)
if with_nms:
det_bboxes, det_labels = multiclass_nms_rotated(
mlvl_bboxes, mlvl_scores, cfg.score_thr, cfg.nms,
cfg.max_per_img)
return det_bboxes, det_labels
else:
raise NotImplementedError