File size: 10,987 Bytes
a277bb8 |
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 |
# ------------------------------------------------------------------------
# DINO
# Copyright (c) 2022 IDEA. All Rights Reserved.
# Licensed under the Apache License, Version 2.0 [see LICENSE for details]
# ------------------------------------------------------------------------
# Modules to compute the matching cost and solve the corresponding LSAP.
# Copyright (c) 2021 Microsoft. All Rights Reserved.
# Licensed under the Apache License, Version 2.0 [see LICENSE for details]
# ------------------------------------------------------------------------
# Modified from DETR (https://github.com/facebookresearch/detr)
# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
# ------------------------------------------------------------------------
# Modified from Deformable DETR (https://github.com/fundamentalvision/Deformable-DETR)
# Copyright (c) 2020 SenseTime. All Rights Reserved.
# ------------------------------------------------------------------------
import torch, os
from torch import nn
from scipy.optimize import linear_sum_assignment
from util.box_ops import box_cxcywh_to_xyxy, generalized_box_iou
class HungarianMatcher(nn.Module):
"""This class computes an assignment between the targets and the predictions of the network
For efficiency reasons, the targets don't include the no_object. Because of this, in general,
there are more predictions than targets. In this case, we do a 1-to-1 matching of the best predictions,
while the others are un-matched (and thus treated as non-objects).
"""
def __init__(self, cost_class: float = 1, cost_bbox: float = 1, cost_giou: float = 1, focal_alpha = 0.25):
"""Creates the matcher
Params:
cost_class: This is the relative weight of the classification error in the matching cost
cost_bbox: This is the relative weight of the L1 error of the bounding box coordinates in the matching cost
cost_giou: This is the relative weight of the giou loss of the bounding box in the matching cost
"""
super().__init__()
self.cost_class = cost_class
self.cost_bbox = cost_bbox
self.cost_giou = cost_giou
assert cost_class != 0 or cost_bbox != 0 or cost_giou != 0, "all costs cant be 0"
self.focal_alpha = focal_alpha
@torch.no_grad()
def forward(self, outputs, targets, label_map):
""" Performs the matching
Params:
outputs: This is a dict that contains at least these entries:
"pred_logits": Tensor of dim [batch_size, num_queries, num_classes] with the classification logits
"pred_boxes": Tensor of dim [batch_size, num_queries, 4] with the predicted box coordinates
targets: This is a list of targets (len(targets) = batch_size), where each target is a dict containing:
"labels": Tensor of dim [num_target_boxes] (where num_target_boxes is the number of ground-truth
objects in the target) containing the class labels
"boxes": Tensor of dim [num_target_boxes, 4] containing the target box coordinates
Returns:
A list of size batch_size, containing tuples of (index_i, index_j) where:
- index_i is the indices of the selected predictions (in order)
- index_j is the indices of the corresponding selected targets (in order)
For each batch element, it holds:
len(index_i) = len(index_j) = min(num_queries, num_target_boxes)
"""
bs, num_queries = outputs["pred_logits"].shape[:2]
# We flatten to compute the cost matrices in a batch
out_prob = outputs["pred_logits"].flatten(0, 1).sigmoid() # [batch_size * num_queries, num_classes]
out_bbox = outputs["pred_boxes"].flatten(0, 1) # [batch_size * num_queries, 4]
# Also concat the target labels and boxes
tgt_ids = torch.cat([v["labels"] for v in targets])
tgt_bbox = torch.cat([v["boxes"] for v in targets])
# Compute the classification cost.
alpha = self.focal_alpha
gamma = 2.0
new_label_map=label_map[tgt_ids.cpu()]
neg_cost_class = (1 - alpha) * (out_prob ** gamma) * (-(1 - out_prob + 1e-8).log())
pos_cost_class = alpha * ((1 - out_prob) ** gamma) * (-(out_prob + 1e-8).log())
new_label_map=new_label_map.to(pos_cost_class.device)
cost_bbox = torch.cdist(out_bbox[:, :2], tgt_bbox[:, :2], p=1)
# cost_class=(pos_cost_class @ new_label_map.T - neg_cost_class@ new_label_map.T)
cost_class=[]
for idx_map in new_label_map:
idx_map = idx_map / idx_map.sum()
cost_class.append(pos_cost_class @ idx_map - neg_cost_class@ idx_map)
if cost_class:
cost_class=torch.stack(cost_class,dim=0).T
else:
cost_class=torch.zeros_like(cost_bbox)
# Compute the L1 cost between boxes
# Compute the giou cost betwen boxes
cost_giou = -generalized_box_iou(box_cxcywh_to_xyxy(out_bbox), box_cxcywh_to_xyxy(tgt_bbox))
# import pdb;pdb.set_trace()
# Final cost matrix
C = self.cost_bbox * cost_bbox + self.cost_class * cost_class + self.cost_giou * cost_giou
C = C.view(bs, num_queries, -1).cpu()
C[torch.isnan(C)] = 0.0
C[torch.isinf(C)] = 0.0
sizes = [len(v["boxes"]) for v in targets]
try:
indices = [linear_sum_assignment(c[i]) for i, c in enumerate(C.split(sizes, -1))]
except:
print("warning: use SimpleMinsumMatcher")
indices = []
device = C.device
for i, (c, _size) in enumerate(zip(C.split(sizes, -1), sizes)):
weight_mat = c[i]
idx_i = weight_mat.min(0)[1]
idx_j = torch.arange(_size).to(device)
indices.append((idx_i, idx_j))
return [(torch.as_tensor(i, dtype=torch.int64), torch.as_tensor(j, dtype=torch.int64)) for i, j in indices]
class SimpleMinsumMatcher(nn.Module):
"""This class computes an assignment between the targets and the predictions of the network
For efficiency reasons, the targets don't include the no_object. Because of this, in general,
there are more predictions than targets. In this case, we do a 1-to-1 matching of the best predictions,
while the others are un-matched (and thus treated as non-objects).
"""
def __init__(self, cost_class: float = 1, cost_bbox: float = 1, cost_giou: float = 1, focal_alpha = 0.25):
"""Creates the matcher
Params:
cost_class: This is the relative weight of the classification error in the matching cost
cost_bbox: This is the relative weight of the L1 error of the bounding box coordinates in the matching cost
cost_giou: This is the relative weight of the giou loss of the bounding box in the matching cost
"""
super().__init__()
self.cost_class = cost_class
self.cost_bbox = cost_bbox
self.cost_giou = cost_giou
assert cost_class != 0 or cost_bbox != 0 or cost_giou != 0, "all costs cant be 0"
self.focal_alpha = focal_alpha
@torch.no_grad()
def forward(self, outputs, targets):
""" Performs the matching
Params:
outputs: This is a dict that contains at least these entries:
"pred_logits": Tensor of dim [batch_size, num_queries, num_classes] with the classification logits
"pred_boxes": Tensor of dim [batch_size, num_queries, 4] with the predicted box coordinates
targets: This is a list of targets (len(targets) = batch_size), where each target is a dict containing:
"labels": Tensor of dim [num_target_boxes] (where num_target_boxes is the number of ground-truth
objects in the target) containing the class labels
"boxes": Tensor of dim [num_target_boxes, 4] containing the target box coordinates
Returns:
A list of size batch_size, containing tuples of (index_i, index_j) where:
- index_i is the indices of the selected predictions (in order)
- index_j is the indices of the corresponding selected targets (in order)
For each batch element, it holds:
len(index_i) = len(index_j) = min(num_queries, num_target_boxes)
"""
bs, num_queries = outputs["pred_logits"].shape[:2]
# We flatten to compute the cost matrices in a batch
out_prob = outputs["pred_logits"].flatten(0, 1).sigmoid() # [batch_size * num_queries, num_classes]
out_bbox = outputs["pred_boxes"].flatten(0, 1) # [batch_size * num_queries, 4]
# Also concat the target labels and boxes
tgt_ids = torch.cat([v["labels"] for v in targets])
tgt_bbox = torch.cat([v["boxes"] for v in targets])
# Compute the classification cost.
alpha = self.focal_alpha
gamma = 2.0
neg_cost_class = (1 - alpha) * (out_prob ** gamma) * (-(1 - out_prob + 1e-8).log())
pos_cost_class = alpha * ((1 - out_prob) ** gamma) * (-(out_prob + 1e-8).log())
cost_class = pos_cost_class[:, tgt_ids] - neg_cost_class[:, tgt_ids]
# Compute the L1 cost between boxes
cost_bbox = torch.cdist(out_bbox, tgt_bbox, p=1)
# Compute the giou cost betwen boxes
cost_giou = -generalized_box_iou(box_cxcywh_to_xyxy(out_bbox), box_cxcywh_to_xyxy(tgt_bbox))
# Final cost matrix
C = self.cost_bbox * cost_bbox + self.cost_class * cost_class + self.cost_giou * cost_giou
C = C.view(bs, num_queries, -1)
sizes = [len(v["boxes"]) for v in targets]
indices = []
device = C.device
for i, (c, _size) in enumerate(zip(C.split(sizes, -1), sizes)):
weight_mat = c[i]
idx_i = weight_mat.min(0)[1]
idx_j = torch.arange(_size).to(device)
indices.append((idx_i, idx_j))
return [(torch.as_tensor(i, dtype=torch.int64), torch.as_tensor(j, dtype=torch.int64)) for i, j in indices]
def build_matcher(args):
assert args.matcher_type in ['HungarianMatcher', 'SimpleMinsumMatcher'], "Unknown args.matcher_type: {}".format(args.matcher_type)
if args.matcher_type == 'HungarianMatcher':
return HungarianMatcher(
cost_class=args.set_cost_class, cost_bbox=args.set_cost_bbox, cost_giou=args.set_cost_giou,
focal_alpha=args.focal_alpha
)
elif args.matcher_type == 'SimpleMinsumMatcher':
return SimpleMinsumMatcher(
cost_class=args.set_cost_class, cost_bbox=args.set_cost_bbox, cost_giou=args.set_cost_giou,
focal_alpha=args.focal_alpha
)
else:
raise NotImplementedError("Unknown args.matcher_type: {}".format(args.matcher_type))
|