Hugging Face's logo

Datasets:
TontonTremblay
/
DOPER_BOP

Tags:
pose-estimation
bop
doper
keypoints
6dof
License:
Dataset card Files
xet
 Community
TontonTremblay commited on
Commit
06e49bc
·
verified ·
1 Parent(s): a0d6c2d

Update README with full pipeline results, all BOP metrics, per-object tables

Browse files
Files changed (1) hide show
  1. README.md +339 -260
README.md CHANGED
@@ -5,288 +5,367 @@ tags:
5
  - bop
6
  - doper
7
  - keypoints
 
8
  ---
9
 
10
- # DOPER BOP -- DOPER-t Keypoint Models for BOP Datasets
11
 
12
- Per-object DOPER-t models trained on synthetic PBR data with symmetry-aware 3D keypoints.
13
- Evaluated using GT bounding boxes, DOPER-t keypoint inference, and PnP pose estimation.
14
 
15
- ## Summary
 
16
 
17
- | Dataset | Objects | Split | GT Instances | PnP Solved | ADD-AUC | ADDS-AUC | MSSD-AUC |
18
- |---------|---------|-------|-------------|------------|---------|----------|----------|
19
- | **lm** | 15 | test | 3000 | 3000 | 0.9812 | 0.9917 | 0.9731 |
20
- | **lmo** | 8 | test | 1517 | 1515 | 0.9680 | 0.9818 | 0.9592 |
21
- | **tless** | 30 | test_primesense | 6900 | 6900 | 0.6807 | 0.7018 | 0.6658 |
22
- | **tudl** | 3 | test | 600 | 600 | 0.9670 | 0.9822 | 0.9486 |
23
- | **ycbv** | 21 | test | 4125 | 4125 | 0.8762 | 0.8975 | 0.8611 |
24
- | **hb** | 33 | val_primesense | 23120 | 23120 | 0.9551 | 0.9640 | 0.9466 |
25
- | **itodd** | 28 | val | 123 | 123 | 0.6588 | 0.6797 | 0.6363 |
26
- | **icbin** | 2 | test | 2250 | 2250 | 0.8250 | 0.8522 | 0.7971 |
27
- | **hope** | 28 | val | 920 | 920 | 0.8813 | 0.8999 | 0.8651 |
28
- | **Overall** | **168** | | **42555** | **42553** | **0.8362** | **0.8531** | **0.8217** |
29
 
30
- > **hb** evaluated on val_primesense (13 scenes). **itodd** evaluated on val (1 scene, grayscale converted to RGB). **hope** evaluated on val (10 scenes).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
 
32
  ## Method
33
 
34
- 1. Generate symmetry-aware 3D keypoints (17 per object) from BOP model meshes + `models_info.json`
35
- 2. Project keypoints onto synthetic PBR training images using GT poses
36
- 3. Train DOPER-t (CSPNeXt-tiny backbone, 256x256 input, 300 epochs, 4x L40S GPUs)
37
- 4. At inference: detect 2D keypoints with DOPER-t, solve 6DoF pose via PnP+RANSAC
38
-
39
- ### Training data
40
-
41
- - **BOP core datasets** (lm, lmo, tless, tudl, ycbv, hb, itodd, icbin): BOP BlenderProc PBR (`train_pbr`)
42
- - **HOPE**: Custom BlenderProc synthetic renders (~40K images per object)
43
-
44
- ### Symmetry-aware keypoints
45
-
46
- | Type | Strategy | Symbol |
47
- |------|----------|--------|
48
- | None | Farthest-point sampling (FPS) on mesh | -- |
49
- | Discrete | Keypoints in fundamental domain, replicated under symmetry transforms | D |
50
- | Continuous | Axial keypoints + equidistant ring perpendicular to symmetry axis | C |
51
-
52
- ## LM (15 objects, eval: test)
53
-
54
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
55
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
56
- | 1 | -- | 102.1 | 200 | 200 | 0.9992 | 1.0000 | 0.9989 | 0.640 | 0.945 | 280 |
57
- | 2 | -- | 247.5 | 200 | 200 | 0.9939 | 0.9991 | 0.9829 | 0.710 | 0.960 | 200 |
58
- | 3 | C | 167.4 | 200 | 200 | 0.9000 | 0.9458 | 0.9046 | 0.005 | 0.340 | 290 |
59
- | 4 | -- | 172.5 | 200 | 200 | 0.9946 | 0.9992 | 0.9935 | 0.485 | 0.785 | 260 |
60
- | 5 | -- | 201.4 | 200 | 200 | 0.9962 | 0.9985 | 0.9941 | 0.875 | 0.975 | 270 |
61
- | 6 | -- | 154.5 | 200 | 200 | 1.0000 | 1.0000 | 1.0000 | 0.910 | 1.000 | 290 |
62
- | 7 | -- | 124.3 | 200 | 200 | 0.9247 | 0.9524 | 0.8945 | 0.075 | 0.440 | 280 |
63
- | 8 | -- | 261.5 | 200 | 200 | 0.9984 | 0.9996 | 0.9950 | 0.905 | 0.960 | 260 |
64
- | 9 | -- | 109.0 | 200 | 200 | 0.9769 | 0.9964 | 0.9560 | 0.140 | 0.735 | 290 |
65
- | 10 | D | 164.6 | 200 | 200 | 0.9929 | 0.9988 | 0.9896 | 0.440 | 0.875 | 300 |
66
- | 11 | D | 175.9 | 200 | 200 | 0.9870 | 0.9964 | 0.9746 | 0.235 | 0.645 | 290 |
67
- | 12 | -- | 145.5 | 200 | 200 | 0.9700 | 0.9907 | 0.9491 | 0.125 | 0.745 | 290 |
68
- | 13 | -- | 278.1 | 200 | 200 | 0.9986 | 0.9995 | 0.9965 | 0.910 | 0.985 | 280 |
69
- | 14 | -- | 282.6 | 200 | 200 | 0.9889 | 0.9999 | 0.9705 | 0.445 | 0.740 | 270 |
70
- | 15 | -- | 212.4 | 200 | 200 | 0.9972 | 1.0000 | 0.9960 | 0.585 | 0.905 | 280 |
71
- | **mean** | | | | | **0.9812** | **0.9917** | **0.9731** | | | |
72
-
73
- ## LMO (8 objects, eval: test)
74
-
75
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
76
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
77
- | 1 | -- | 102.1 | 187 | 187 | 0.9852 | 0.9901 | 0.9806 | 0.417 | 0.781 | 250 |
78
- | 5 | -- | 201.4 | 199 | 199 | 0.9941 | 0.9989 | 0.9879 | 0.709 | 0.920 | 300 |
79
- | 6 | -- | 154.5 | 196 | 196 | 0.9855 | 0.9899 | 0.9807 | 0.423 | 0.765 | 250 |
80
- | 8 | -- | 261.5 | 200 | 200 | 0.9980 | 0.9999 | 0.9936 | 0.800 | 0.955 | 250 |
81
- | 9 | -- | 109.0 | 188 | 188 | 0.9608 | 0.9814 | 0.9379 | 0.069 | 0.649 | 300 |
82
- | 10 | D | 164.6 | 193 | 191 | 0.8827 | 0.9247 | 0.8868 | 0.005 | 0.330 | 290 |
83
- | 11 | D | 175.9 | 154 | 154 | 0.9666 | 0.9765 | 0.9544 | 0.305 | 0.571 | 270 |
84
- | 12 | -- | 145.5 | 200 | 200 | 0.9715 | 0.9931 | 0.9516 | 0.105 | 0.750 | 270 |
85
- | **mean** | | | | | **0.9680** | **0.9818** | **0.9592** | | | |
86
-
87
- ## TLESS (30 objects, eval: test_primesense)
88
-
89
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
90
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
91
- | 1 | C | 63.5 | 900 | 900 | 0.7896 | 0.8190 | 0.7684 | 0.006 | 0.182 | 300 |
92
- | 2 | C | 66.2 | 500 | 500 | 0.0641 | 0.0722 | 0.0601 | 0.000 | 0.000 | 170 |
93
- | 3 | C | 65.3 | 400 | 400 | 0.8174 | 0.8523 | 0.7965 | 0.010 | 0.212 | 300 |
94
- | 4 | C | 80.7 | 650 | 650 | 0.8842 | 0.9095 | 0.8655 | 0.038 | 0.262 | 300 |
95
- | 5 | D | 108.7 | 200 | 200 | 0.9652 | 0.9779 | 0.9589 | 0.345 | 0.640 | 260 |
96
- | 6 | D | 108.3 | 100 | 100 | 0.9215 | 0.9408 | 0.9115 | 0.120 | 0.280 | 290 |
97
- | 7 | D | 178.6 | 250 | 250 | 0.8148 | 0.8378 | 0.7952 | 0.244 | 0.400 | 300 |
98
- | 8 | D | 217.2 | 150 | 150 | 0.9193 | 0.9367 | 0.9078 | 0.073 | 0.420 | 290 |
99
- | 9 | D | 144.5 | 250 | 250 | 0.8726 | 0.9124 | 0.8329 | 0.008 | 0.132 | 280 |
100
- | 10 | D | 90.2 | 150 | 150 | 0.9510 | 0.9635 | 0.9433 | 0.393 | 0.720 | 300 |
101
- | 11 | D | 76.6 | 200 | 200 | 0.9255 | 0.9459 | 0.9122 | 0.155 | 0.595 | 300 |
102
- | 12 | D | 86.0 | 150 | 150 | 0.9265 | 0.9478 | 0.9103 | 0.267 | 0.680 | 290 |
103
- | 13 | C | 58.1 | 150 | 150 | 0.8528 | 0.8860 | 0.8363 | 0.040 | 0.247 | 300 |
104
- | 14 | C | 71.9 | 150 | 150 | 0.2685 | 0.2898 | 0.2555 | 0.000 | 0.000 | 260 |
105
- | 15 | C | 68.6 | 150 | 150 | 0.8388 | 0.8735 | 0.8162 | 0.027 | 0.187 | 300 |
106
- | 16 | C | 69.2 | 200 | 200 | 0.0186 | 0.0220 | 0.0167 | 0.000 | 0.000 | 250 |
107
- | 17 | C | 112.8 | 150 | 150 | 0.1103 | 0.1243 | 0.1052 | 0.000 | 0.007 | 300 |
108
- | 18 | -- | 111.0 | 150 | 150 | 0.9632 | 0.9803 | 0.9412 | 0.340 | 0.740 | 300 |
109
- | 19 | D | 89.1 | 200 | 200 | 0.0006 | 0.0012 | 0.0001 | 0.000 | 0.000 | 300 |
110
- | 20 | D | 98.9 | 250 | 250 | 0.4118 | 0.4508 | 0.3812 | 0.000 | 0.000 | 240 |
111
- | 21 | -- | 92.3 | 200 | 200 | 0.9085 | 0.9269 | 0.8896 | 0.420 | 0.610 | 300 |
112
- | 22 | -- | 92.3 | 200 | 200 | 0.9520 | 0.9654 | 0.9379 | 0.475 | 0.655 | 280 |
113
- | 23 | D | 142.6 | 250 | 250 | 0.9738 | 0.9852 | 0.9688 | 0.476 | 0.764 | 290 |
114
- | 24 | C | 84.7 | 200 | 200 | 0.9224 | 0.9470 | 0.9116 | 0.175 | 0.525 | 300 |
115
- | 25 | D | 108.8 | 100 | 100 | 0.9387 | 0.9477 | 0.9328 | 0.640 | 0.760 | 300 |
116
- | 26 | D | 108.8 | 100 | 100 | 0.9062 | 0.9320 | 0.8722 | 0.000 | 0.220 | 40 |
117
- | 27 | D | 152.5 | 100 | 100 | 0.2262 | 0.2485 | 0.2105 | 0.000 | 0.000 | 80 |
118
- | 28 | D | 124.8 | 200 | 200 | 0.2851 | 0.3164 | 0.2576 | 0.000 | 0.000 | 300 |
119
- | 29 | D | 134.2 | 100 | 100 | 0.8903 | 0.9300 | 0.8865 | 0.110 | 0.290 | 300 |
120
- | 30 | C | 88.8 | 150 | 150 | 0.1000 | 0.1125 | 0.0912 | 0.000 | 0.000 | 220 |
121
- | **mean** | | | | | **0.6807** | **0.7018** | **0.6658** | | | |
122
-
123
- ## TUDL (3 objects, eval: test)
124
-
125
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
126
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
127
- | 1 | -- | 430.3 | 200 | 200 | 0.9715 | 0.9823 | 0.9597 | 0.265 | 0.560 | 240 |
128
- | 2 | -- | 175.7 | 200 | 200 | 0.9442 | 0.9667 | 0.9238 | 0.115 | 0.270 | 160 |
129
- | 3 | -- | 352.4 | 200 | 200 | 0.9852 | 0.9975 | 0.9624 | 0.290 | 0.685 | 160 |
130
- | **mean** | | | | | **0.9670** | **0.9822** | **0.9486** | | | |
131
-
132
- ## YCBV (21 objects, eval: test)
133
-
134
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
135
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
136
- | 1 | D | 172.1 | 300 | 300 | 0.9496 | 0.9830 | 0.9501 | 0.000 | 0.027 | 200 |
137
- | 2 | -- | 269.6 | 225 | 225 | 0.9970 | 0.9997 | 0.9956 | 0.231 | 0.698 | 210 |
138
- | 3 | -- | 198.4 | 375 | 375 | 0.9950 | 0.9991 | 0.9935 | 0.331 | 0.723 | 250 |
139
- | 4 | -- | 120.5 | 450 | 450 | 0.9606 | 0.9740 | 0.9531 | 0.189 | 0.391 | 280 |
140
- | 5 | -- | 196.5 | 150 | 150 | 0.9495 | 0.9633 | 0.9357 | 0.013 | 0.247 | 290 |
141
- | 6 | -- | 89.8 | 300 | 300 | 0.7738 | 0.8163 | 0.7187 | 0.000 | 0.060 | 300 |
142
- | 7 | -- | 142.5 | 75 | 75 | 0.9820 | 0.9933 | 0.9770 | 0.080 | 0.267 | 280 |
143
- | 8 | -- | 114.1 | 75 | 75 | 0.9907 | 1.0000 | 0.9860 | 0.173 | 0.387 | 280 |
144
- | 9 | -- | 129.5 | 225 | 225 | 0.8667 | 0.8942 | 0.8384 | 0.142 | 0.284 | 270 |
145
- | 10 | -- | 197.8 | 150 | 150 | 0.7397 | 0.7570 | 0.7125 | 0.000 | 0.013 | 290 |
146
- | 11 | -- | 259.5 | 225 | 225 | 0.7047 | 0.7330 | 0.6739 | 0.000 | 0.000 | 230 |
147
- | 12 | -- | 259.6 | 300 | 300 | 0.7275 | 0.7411 | 0.7122 | 0.083 | 0.217 | 270 |
148
- | 13 | C | 161.9 | 150 | 150 | 0.9005 | 0.9490 | 0.8928 | 0.000 | 0.080 | 290 |
149
- | 14 | -- | 125.0 | 150 | 150 | 0.9588 | 0.9855 | 0.9280 | 0.013 | 0.140 | 270 |
150
- | 15 | -- | 226.2 | 300 | 300 | 0.9940 | 0.9968 | 0.9889 | 0.103 | 0.687 | 220 |
151
- | 16 | D | 237.3 | 75 | 75 | 0.6937 | 0.7260 | 0.6763 | 0.000 | 0.000 | 300 |
152
- | 17 | -- | 204.0 | 75 | 75 | 0.9517 | 0.9683 | 0.9507 | 0.000 | 0.000 | 280 |
153
- | 18 | C | 121.4 | 150 | 150 | 0.9842 | 0.9907 | 0.9812 | 0.220 | 0.613 | 230 |
154
- | 19 | D | 174.7 | 150 | 150 | 0.9267 | 0.9523 | 0.9173 | 0.000 | 0.040 | 290 |
155
- | 20 | D | 217.1 | 150 | 150 | 0.9218 | 0.9550 | 0.9058 | 0.007 | 0.247 | 300 |
156
- | 21 | D | 102.9 | 75 | 75 | 0.4330 | 0.4703 | 0.3950 | 0.000 | 0.000 | 290 |
157
- | **mean** | | | | | **0.8762** | **0.8975** | **0.8611** | | | |
158
-
159
- ## HB (33 objects, eval: val_primesense)
160
-
161
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
162
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
163
- | 1 | -- | 232.6 | 680 | 680 | 0.9912 | 0.9967 | 0.9813 | 0.568 | 0.850 | 290 |
164
- | 2 | -- | 257.4 | 340 | 340 | 1.0000 | 1.0000 | 1.0000 | 1.000 | 1.000 | 270 |
165
- | 3 | -- | 166.5 | 1020 | 1020 | 0.9745 | 0.9830 | 0.9656 | 0.659 | 0.818 | 260 |
166
- | 4 | -- | 179.0 | 680 | 680 | 0.9856 | 0.9911 | 0.9795 | 0.566 | 0.785 | 290 |
167
- | 5 | -- | 205.4 | 680 | 680 | 0.9873 | 0.9908 | 0.9835 | 0.776 | 0.838 | 300 |
168
- | 6 | -- | 121.4 | 340 | 340 | 0.9746 | 0.9971 | 0.9477 | 0.097 | 0.724 | 300 |
169
- | 7 | -- | 263.7 | 340 | 340 | 0.9987 | 0.9999 | 0.9949 | 0.938 | 0.959 | 290 |
170
- | 8 | -- | 186.8 | 680 | 680 | 0.9953 | 0.9979 | 0.9929 | 0.754 | 0.904 | 300 |
171
- | 9 | -- | 166.6 | 680 | 680 | 0.9200 | 0.9401 | 0.8964 | 0.341 | 0.601 | 290 |
172
- | 10 | C | 180.8 | 680 | 680 | 0.9365 | 0.9575 | 0.9327 | 0.410 | 0.601 | 300 |
173
- | 11 | C | 238.5 | 340 | 340 | 0.9277 | 0.9538 | 0.9388 | 0.526 | 0.715 | 280 |
174
- | 12 | -- | 156.9 | 1360 | 1360 | 0.9745 | 0.9868 | 0.9604 | 0.348 | 0.724 | 300 |
175
- | 13 | -- | 145.3 | 1020 | 1020 | 0.9574 | 0.9717 | 0.9466 | 0.147 | 0.397 | 290 |
176
- | 14 | C | 243.7 | 680 | 680 | 0.9922 | 0.9968 | 0.9868 | 0.688 | 0.854 | 280 |
177
- | 15 | -- | 113.0 | 1700 | 1700 | 0.9502 | 0.9629 | 0.9405 | 0.314 | 0.571 | 280 |
178
- | 16 | -- | 101.6 | 1020 | 1020 | 0.9588 | 0.9743 | 0.9376 | 0.227 | 0.465 | 280 |
179
- | 17 | -- | 132.8 | 1360 | 1360 | 0.9460 | 0.9579 | 0.9306 | 0.335 | 0.604 | 290 |
180
- | 18 | -- | 211.1 | 680 | 680 | 0.1994 | 0.2258 | 0.1693 | 0.000 | 0.000 | 270 |
181
- | 19 | -- | 185.6 | 680 | 680 | 0.9987 | 0.9993 | 0.9982 | 0.949 | 0.976 | 280 |
182
- | 20 | -- | 244.8 | 340 | 340 | 1.0000 | 1.0000 | 1.0000 | 0.976 | 1.000 | 270 |
183
- | 21 | -- | 212.6 | 340 | 340 | 0.9979 | 0.9993 | 0.9963 | 0.812 | 0.921 | 300 |
184
- | 22 | -- | 190.2 | 1360 | 1360 | 0.9874 | 0.9933 | 0.9823 | 0.551 | 0.780 | 270 |
185
- | 23 | -- | 233.9 | 1020 | 1020 | 0.9987 | 0.9997 | 0.9971 | 0.801 | 0.941 | 280 |
186
- | 24 | -- | 252.3 | 340 | 340 | 0.9940 | 0.9999 | 0.9889 | 0.632 | 0.926 | 300 |
187
- | 25 | -- | 202.9 | 680 | 680 | 0.9753 | 0.9853 | 0.9612 | 0.409 | 0.631 | 290 |
188
- | 26 | -- | 183.8 | 680 | 680 | 0.9785 | 0.9893 | 0.9613 | 0.443 | 0.718 | 290 |
189
- | 27 | -- | 264.4 | 340 | 340 | 0.9989 | 0.9998 | 0.9983 | 0.821 | 0.985 | 300 |
190
- | 28 | -- | 477.5 | 340 | 340 | 1.0000 | 1.0000 | 1.0000 | 1.000 | 1.000 | 280 |
191
- | 29 | -- | 198.0 | 680 | 680 | 0.9720 | 0.9970 | 0.9489 | 0.496 | 0.851 | 300 |
192
- | 30 | -- | 416.2 | 340 | 340 | 1.0000 | 1.0000 | 1.0000 | 0.982 | 1.000 | 240 |
193
- | 31 | -- | 158.0 | 340 | 340 | 0.9912 | 0.9971 | 0.9817 | 0.526 | 0.800 | 280 |
194
- | 32 | -- | 201.8 | 680 | 680 | 0.9610 | 0.9701 | 0.9487 | 0.481 | 0.696 | 290 |
195
- | 33 | -- | 187.2 | 680 | 680 | 0.9945 | 0.9974 | 0.9916 | 0.788 | 0.906 | 300 |
196
- | **mean** | | | | | **0.9551** | **0.9640** | **0.9466** | | | |
197
-
198
- ## ITODD (28 objects, eval: val)
199
-
200
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
201
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
202
- | 1 | -- | 64.1 | 4 | 4 | 0.8875 | 0.9250 | 0.8375 | 0.000 | 0.250 | 300 |
203
- | 2 | D | 51.5 | 3 | 3 | 0.9417 | 0.9667 | 0.8917 | 0.000 | 0.000 | 120 |
204
- | 3 | D | 142.2 | 4 | 4 | 0.0000 | 0.0000 | 0.0000 | 0.000 | 0.000 | 200 |
205
- | 4 | D | 139.4 | 3 | 3 | 0.3667 | 0.4000 | 0.3500 | 0.000 | 0.000 | 260 |
206
- | 5 | D | 158.6 | 6 | 6 | 0.9125 | 0.9500 | 0.8792 | 0.000 | 0.000 | 230 |
207
- | 6 | -- | 85.3 | 5 | 5 | 0.9650 | 0.9750 | 0.9550 | 0.000 | 0.000 | 280 |
208
- | 7 | C | 38.5 | 5 | 5 | 0.2400 | 0.2650 | 0.2200 | 0.000 | 0.000 | 30 |
209
- | 8 | D | 68.9 | 3 | 3 | 0.8083 | 0.8500 | 0.7417 | 0.000 | 0.000 | 130 |
210
- | 9 | D | 94.8 | 3 | 3 | 0.0000 | 0.0000 | 0.0000 | 0.000 | 0.000 | 90 |
211
- | 10 | -- | 55.7 | 4 | 4 | 0.9187 | 0.9375 | 0.8875 | 0.000 | 0.000 | 30 |
212
- | 11 | D | 140.1 | 5 | 5 | 0.6450 | 0.6750 | 0.6300 | 0.000 | 0.000 | 230 |
213
- | 12 | D | 107.7 | 4 | 4 | 0.0000 | 0.0000 | 0.0000 | 0.000 | 0.000 | 30 |
214
- | 13 | -- | 128.1 | 4 | 4 | 0.9812 | 0.9938 | 0.9688 | 0.250 | 0.250 | 290 |
215
- | 14 | D | 102.9 | 3 | 3 | 0.0000 | 0.0000 | 0.0000 | 0.000 | 0.000 | 290 |
216
- | 15 | -- | 114.2 | 3 | 3 | 0.9833 | 1.0000 | 0.9750 | 0.333 | 1.000 | 250 |
217
- | 16 | -- | 193.1 | 3 | 3 | 0.9917 | 1.0000 | 0.9750 | 0.333 | 0.333 | 290 |
218
- | 17 | D | 77.8 | 3 | 3 | 0.7667 | 0.8167 | 0.7000 | 0.000 | 0.000 | 300 |
219
- | 18 | D | 108.5 | 3 | 3 | 0.2000 | 0.2333 | 0.1833 | 0.000 | 0.000 | 300 |
220
- | 19 | D | 121.4 | 3 | 3 | 0.5333 | 0.5667 | 0.5000 | 0.000 | 0.000 | 300 |
221
- | 20 | -- | 122.0 | 4 | 4 | 0.9812 | 0.9875 | 0.9688 | 0.250 | 0.500 | 280 |
222
- | 21 | -- | 171.2 | 3 | 3 | 1.0000 | 1.0000 | 1.0000 | 1.000 | 1.000 | 300 |
223
- | 22 | -- | 267.5 | 3 | 3 | 0.9250 | 0.9583 | 0.8750 | 0.000 | 0.000 | 300 |
224
- | 23 | D | 56.9 | 1 | 1 | 1.0000 | 1.0000 | 1.0000 | 0.000 | 1.000 | 300 |
225
- | 24 | C | 65.0 | 6 | 6 | 0.3167 | 0.3292 | 0.3083 | 0.000 | 0.000 | 20 |
226
- | 25 | D | 48.5 | 6 | 6 | 0.9500 | 0.9750 | 0.9125 | 0.000 | 0.000 | 280 |
227
- | 26 | -- | 66.8 | 6 | 6 | 0.9792 | 0.9958 | 0.9667 | 0.167 | 0.500 | 290 |
228
- | 27 | C | 55.7 | 5 | 5 | 0.7600 | 0.8100 | 0.7150 | 0.000 | 0.000 | 270 |
229
- | 28 | D | 24.1 | 18 | 18 | 0.3917 | 0.4222 | 0.3764 | 0.000 | 0.000 | 150 |
230
- | **mean** | | | | | **0.6588** | **0.6797** | **0.6363** | | | |
231
-
232
- ## ICBIN (2 objects, eval: test)
233
-
234
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
235
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
236
- | 1 | D | 136.6 | 1800 | 1800 | 0.8103 | 0.8420 | 0.7911 | 0.078 | 0.226 | 30 |
237
- | 2 | -- | 220.6 | 450 | 450 | 0.8396 | 0.8624 | 0.8031 | 0.011 | 0.042 | 30 |
238
- | **mean** | | | | | **0.8250** | **0.8522** | **0.7971** | | | |
239
-
240
- ## HOPE (28 objects, eval: val)
241
-
242
- | ID | Sym | Diam (mm) | N GT | Solved | ADD-AUC | ADDS-AUC | MSSD-AUC | ADD<10% | ADDS<10% | Epoch |
243
- |----|-----|-----------|------|--------|---------|----------|----------|---------|----------|-------|
244
- | 1 | -- | 107.9 | 45 | 45 | 0.9017 | 0.9172 | 0.8928 | 0.267 | 0.400 | 200 |
245
- | 2 | -- | 153.9 | 45 | 45 | 0.9211 | 0.9389 | 0.8956 | 0.178 | 0.333 | 300 |
246
- | 3 | -- | 115.0 | 45 | 45 | 0.8372 | 0.8522 | 0.8239 | 0.356 | 0.422 | 90 |
247
- | 4 | -- | 89.6 | 20 | 20 | 0.9513 | 0.9688 | 0.9412 | 0.100 | 0.400 | 260 |
248
- | 5 | -- | 98.0 | 30 | 30 | 0.8267 | 0.8417 | 0.8125 | 0.367 | 0.433 | 220 |
249
- | 6 | -- | 208.7 | 25 | 25 | 0.9940 | 0.9990 | 0.9920 | 0.320 | 0.720 | 300 |
250
- | 7 | -- | 89.6 | 40 | 40 | 0.8538 | 0.8900 | 0.8300 | 0.075 | 0.175 | 140 |
251
- | 8 | -- | 115.6 | 25 | 25 | 0.9290 | 0.9480 | 0.9080 | 0.200 | 0.360 | 30 |
252
- | 9 | -- | 206.0 | 25 | 25 | 0.8520 | 0.8780 | 0.8300 | 0.080 | 0.360 | 120 |
253
- | 10 | -- | 89.7 | 35 | 35 | 0.5721 | 0.5957 | 0.5471 | 0.114 | 0.314 | 80 |
254
- | 11 | -- | 153.9 | 40 | 40 | 0.9338 | 0.9444 | 0.9244 | 0.325 | 0.575 | 80 |
255
- | 12 | -- | 207.1 | 50 | 50 | 0.9300 | 0.9525 | 0.8950 | 0.220 | 0.420 | 180 |
256
- | 13 | -- | 153.4 | 30 | 30 | 0.9742 | 0.9883 | 0.9617 | 0.200 | 0.333 | 230 |
257
- | 14 | -- | 204.5 | 25 | 25 | 0.9890 | 0.9970 | 0.9790 | 0.160 | 0.520 | 100 |
258
- | 15 | -- | 75.7 | 35 | 35 | 0.8907 | 0.9279 | 0.8679 | 0.000 | 0.029 | 90 |
259
- | 16 | -- | 161.5 | 50 | 50 | 0.9425 | 0.9515 | 0.9300 | 0.440 | 0.640 | 200 |
260
- | 17 | -- | 205.7 | 25 | 25 | 0.7170 | 0.7320 | 0.6780 | 0.160 | 0.280 | 230 |
261
- | 18 | -- | 122.8 | 30 | 30 | 0.9300 | 0.9408 | 0.9192 | 0.500 | 0.700 | 210 |
262
- | 19 | -- | 89.2 | 20 | 20 | 0.9587 | 0.9750 | 0.9550 | 0.150 | 0.300 | 70 |
263
- | 20 | -- | 89.9 | 30 | 30 | 0.7617 | 0.7933 | 0.7433 | 0.100 | 0.100 | 100 |
264
- | 21 | -- | 89.2 | 55 | 55 | 0.8659 | 0.8877 | 0.8509 | 0.182 | 0.327 | 270 |
265
- | 22 | -- | 152.4 | 20 | 20 | 0.8100 | 0.8375 | 0.7900 | 0.000 | 0.050 | 290 |
266
- | 23 | -- | 151.3 | 40 | 40 | 0.8850 | 0.9125 | 0.8525 | 0.075 | 0.225 | 180 |
267
- | 24 | -- | 151.3 | 10 | 10 | 0.9675 | 0.9725 | 0.9525 | 0.400 | 0.400 | 260 |
268
- | 25 | -- | 252.8 | 35 | 35 | 0.9793 | 0.9886 | 0.9736 | 0.486 | 0.571 | 220 |
269
- | 26 | -- | 107.1 | 35 | 35 | 0.7779 | 0.7929 | 0.7664 | 0.286 | 0.429 | 290 |
270
- | 27 | -- | 76.1 | 35 | 35 | 0.7586 | 0.7857 | 0.7514 | 0.000 | 0.143 | 60 |
271
- | 28 | -- | 82.9 | 20 | 20 | 0.9663 | 0.9875 | 0.9587 | 0.100 | 0.250 | 190 |
272
- | **mean** | | | | | **0.8813** | **0.8999** | **0.8651** | | | |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
273
 
274
  ## File Structure
275
 
276
  ```
277
  {dataset}/obj_{NNNNNN}/
278
- best_coco_AP_epoch_NNN.pth # DOPER-t checkpoint
279
  keypoints_3d.json # 17 symmetry-aware 3D keypoints (mm)
280
  bop_summary.json # Evaluation metrics
281
- vis_grid.jpg # Qualitative results
282
  ```
283
 
 
 
284
  ## Usage
285
 
286
  ```python
287
  from huggingface_hub import hf_hub_download
288
 
289
- # Download checkpoint + keypoints for ycbv object 1
290
- ckpt = hf_hub_download("TontonTremblay/DOPER_BOP", "ycbv/obj_000001/best_coco_AP_epoch_280.pth", repo_type="dataset")
291
  kpts = hf_hub_download("TontonTremblay/DOPER_BOP", "ycbv/obj_000001/keypoints_3d.json", repo_type="dataset")
292
  ```
 
 
 
 
 
 
 
 
 
 
 
5
  - bop
6
  - doper
7
  - keypoints
8
+ - 6dof
9
  ---
10
 
11
+ # DOPER BOP -- Per-Object 6DoF Pose Estimation on BOP Datasets
12
 
13
+ Per-object **DOPER-t** (keypoint) and **RTMDet-tiny** (detector) models for all 168 objects across 9 BOP datasets.
14
+ Trained on synthetic PBR data only. Supports two evaluation modes:
15
 
16
+ 1. **Keypoints-only** (GT bounding box) -- upper bound on pose accuracy
17
+ 2. **Detection + Pose** (full pipeline, no GT) -- RTMDet detection, DOPER-t keypoints, PnP solve
18
 
19
+ ## Demo: Full Pipeline (Detection + Keypoints)
 
 
 
 
 
 
 
 
 
 
 
20
 
21
+ ![DOPER-t detection + keypoints on HOPE val](demo/doper_hope_det_keypoints.png)
22
+
23
+ *28 per-object RTMDet detectors + 28 per-object DOPER-t keypoint models running on a single HOPE val image. No GT bounding boxes.*
24
+
25
+ ## Results Summary
26
+
27
+ ### Keypoints-Only (GT BBox)
28
+
29
+ | Dataset | Objects | Split | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
30
+ |---------|---------|-------|------|---------|-----------|----------|----------|-----------|-----------|
31
+ | **lm** | 15 | test | 3000 | 0.9820 | 0.9931 | 0.9733 | 0.9766 | 0.6166 | 0.7718 |
32
+ | **lmo** | 8 | test | 1517 | 0.9655 | 0.9798 | 0.9565 | 0.9786 | 0.5042 | 0.7479 |
33
+ | **tless** | 30 | test_primesense | 6900 | 0.6816 | 0.7030 | 0.6662 | 0.9140 | 0.2245 | 0.4098 |
34
+ | **tudl** | 3 | test | 600 | 0.9635 | 0.9789 | 0.9431 | 0.9587 | 0.3465 | 0.6038 |
35
+ | **ycbv** | 21 | test | 4125 | 0.8671 | 0.8884 | 0.8503 | 0.9177 | 0.2528 | 0.4584 |
36
+ | **hb** | 33 | val_primesense | 23120 | 0.9538 | 0.9630 | 0.9451 | 0.9772 | 0.5961 | 0.7718 |
37
+ | **itodd** | 28 | val | 123 | 0.6352 | 0.6582 | 0.6123 | 0.6237 | 0.0925 | 0.1037 |
38
+ | **icbin** | 2 | test | 2250 | 0.8641 | 0.8921 | 0.8372 | 0.9192 | 0.1548 | 0.3391 |
39
+ | **hope** | 28 | val | 920 | 0.8759 | 0.8937 | 0.8581 | 0.8878 | 0.3331 | 0.4424 |
40
+
41
+ ### Detection + Pose (Full Pipeline)
42
+
43
+ Per-object RTMDet-tiny detectors trained with ~30-40% negative images (cross-dataset).
44
+
45
+ | Dataset | Objects | Split | N GT | Det Rate | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
46
+ |---------|---------|-------|------|----------|---------|-----------|----------|----------|-----------|-----------|
47
+ | **lm** | 15 | test | 3000 | 97.4% | 0.9843 | 0.9950 | 0.9768 | 0.9783 | 0.6264 | 0.7813 |
48
+ | **lmo** | 8 | test | 1517 | 94.5% | 0.9701 | 0.9840 | 0.9622 | 0.9801 | 0.5234 | 0.7602 |
49
+ | **tless** | 30 | test_primesense | 6900 | 74.1% | 0.7079 | 0.7292 | 0.6940 | 0.9129 | 0.2581 | 0.4351 |
50
+ | **tudl** | 3 | test | 600 | 93.8% | 0.9064 | 0.9251 | 0.8831 | 0.9380 | 0.3032 | 0.5281 |
51
+ | **ycbv** | 21 | test | 4125 | 85.0% | 0.7650 | 0.7882 | 0.7473 | 0.8695 | 0.2001 | 0.3671 |
52
+ | **itodd** | 28 | val | 123 | 66.7% | 0.6369 | 0.6554 | 0.6202 | 0.5718 | 0.1609 | 0.1177 |
53
+ | **icbin** | 2 | test | 2250 | 9.1% | 0.9116 | 0.9414 | 0.8826 | 0.8452 | 0.1506 | 0.2443 |
54
+ | **hope** | 28 | val | 920 | 57.8% | 0.9199 | 0.9362 | 0.9065 | 0.8856 | 0.3605 | 0.4596 |
55
+
56
+ > Det Rate = fraction of GT instances matched by a detection (IoU > 0.1). AUC metrics computed only over detected+solved instances.
57
+
58
+ ## Metrics
59
+
60
+ | Metric | Description |
61
+ |--------|-------------|
62
+ | **ADD** | Average Distance of model points (non-symmetric) |
63
+ | **ADD-S** | Average Distance of closest model points (symmetric-aware, a.k.a. ADI) |
64
+ | **MSSD** | Maximum Symmetry-Aware Surface Distance ([BOP](http://bop.felk.cvut.cz/challenges/bop-challenge-2019/)) |
65
+ | **MSPD** | Maximum Symmetry-Aware Projection Distance ([BOP](http://bop.felk.cvut.cz/challenges/bop-challenge-2019/)) |
66
+ | **AUC** | Area Under the recall-vs-threshold Curve (40 thresholds in (0, 10x diameter]) |
67
+ | **AR<sub>MSSD</sub>** | Average Recall at BOP thresholds: {0.05, 0.10, ..., 0.50} x diameter |
68
+ | **AR<sub>MSPD</sub>** | Average Recall at BOP thresholds: {5, 10, ..., 50} pixels |
69
 
70
  ## Method
71
 
72
+ 1. **3D Keypoints**: 17 symmetry-aware keypoints per object from BOP meshes + `models_info.json`
73
+ 2. **Keypoint Training**: DOPER-t (CSPNeXt-tiny, 256x256, 300 epochs) on projected keypoints from PBR renders
74
+ 3. **Detector Training**: RTMDet-tiny (CSPNeXt-tiny, COCO pretrained, 20 epochs) per object, with ~30-40% cross-dataset negative images
75
+ 4. **Inference**: RTMDet detects object bbox -> DOPER-t predicts 2D keypoints -> PnP+RANSAC solves 6DoF pose
76
+
77
+ ### Training Data
78
+
79
+ - **BOP core** (lm, lmo, tless, tudl, ycbv, hb, itodd, icbin): BOP BlenderProc PBR (`train_pbr`)
80
+ - **HOPE**: Custom BlenderProc synthetic (~40K images/object)
81
+
82
+ ### Symmetry Handling
83
+
84
+ | Type | Strategy |
85
+ |------|----------|
86
+ | None | Farthest-point sampling (FPS) on mesh surface |
87
+ | Discrete | Keypoints in fundamental domain, replicated under symmetry transforms |
88
+ | Continuous | Axial keypoints + equidistant ring perpendicular to symmetry axis |
89
+
90
+ ## Per-Object Results (Keypoints-Only, GT BBox)
91
+
92
+ <details>
93
+ <summary><b>LM</b> (15 objects, test)</summary>
94
+
95
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
96
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
97
+ | 1 | 102.1 | 200 | 0.9990 | 1.0000 | 0.9986 | 1.0000 | 0.8260 | 0.9965 |
98
+ | 2 | 247.5 | 200 | 0.9926 | 0.9995 | 0.9779 | 0.9622 | 0.6405 | 0.6985 |
99
+ | 3 | 167.4 | 200 | 0.9106 | 0.9558 | 0.9161 | 0.9319 | 0.2665 | 0.4410 |
100
+ | 4 | 172.5 | 200 | 0.9936 | 0.9990 | 0.9922 | 0.9980 | 0.6900 | 0.9295 |
101
+ | 5 | 201.4 | 200 | 0.9955 | 0.9986 | 0.9915 | 0.9914 | 0.8490 | 0.9080 |
102
+ | 6 | 154.5 | 200 | 1.0000 | 1.0000 | 1.0000 | 1.0000 | 0.9285 | 0.9950 |
103
+ | 7 | 124.3 | 200 | 0.9335 | 0.9620 | 0.9018 | 0.9450 | 0.1750 | 0.4435 |
104
+ | 8 | 261.5 | 200 | 0.9977 | 0.9995 | 0.9936 | 0.9939 | 0.8770 | 0.9205 |
105
+ | 9 | 109.0 | 200 | 0.9749 | 0.9949 | 0.9527 | 0.9672 | 0.3800 | 0.6435 |
106
+ | 10 | 164.6 | 200 | 0.9911 | 0.9975 | 0.9875 | 0.9878 | 0.6850 | 0.8915 |
107
+ | 11 | 175.9 | 200 | 0.9851 | 0.9965 | 0.9730 | 0.9940 | 0.4825 | 0.8785 |
108
+ | 12 | 145.5 | 200 | 0.9734 | 0.9950 | 0.9521 | 0.9456 | 0.3175 | 0.5175 |
109
+ | 13 | 278.1 | 200 | 0.9982 | 0.9994 | 0.9970 | 0.9968 | 0.8965 | 0.9090 |
110
+ | 14 | 282.6 | 200 | 0.9878 | 0.9991 | 0.9695 | 0.9370 | 0.4830 | 0.4875 |
111
+ | 15 | 212.4 | 200 | 0.9972 | 1.0000 | 0.9957 | 0.9989 | 0.7515 | 0.9165 |
112
+ | **mean** | | | **0.9820** | **0.9931** | **0.9733** | **0.9766** | **0.6166** | **0.7718** |
113
+
114
+ </details>
115
+
116
+ <details>
117
+ <summary><b>LMO</b> (8 objects, test)</summary>
118
+
119
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
120
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
121
+ | 1 | 102.1 | 187 | 0.9828 | 0.9880 | 0.9790 | 0.9951 | 0.6567 | 0.8995 |
122
+ | 5 | 201.4 | 199 | 0.9935 | 0.9987 | 0.9861 | 0.9864 | 0.7111 | 0.8201 |
123
+ | 6 | 154.5 | 196 | 0.9858 | 0.9909 | 0.9811 | 0.9971 | 0.6143 | 0.9276 |
124
+ | 8 | 261.5 | 200 | 0.9981 | 0.9996 | 0.9949 | 0.9954 | 0.8350 | 0.8840 |
125
+ | 9 | 109.0 | 188 | 0.9608 | 0.9826 | 0.9368 | 0.9654 | 0.2883 | 0.5968 |
126
+ | 10 | 164.6 | 191 | 0.8719 | 0.9132 | 0.8753 | 0.9677 | 0.2435 | 0.6429 |
127
+ | 11 | 175.9 | 154 | 0.9625 | 0.9735 | 0.9524 | 0.9828 | 0.4422 | 0.7721 |
128
+ | 12 | 145.5 | 200 | 0.9689 | 0.9920 | 0.9468 | 0.9389 | 0.2430 | 0.4405 |
129
+ | **mean** | | | **0.9655** | **0.9798** | **0.9565** | **0.9786** | **0.5042** | **0.7479** |
130
+
131
+ </details>
132
+
133
+ <details>
134
+ <summary><b>T-LESS</b> (30 objects, test_primesense)</summary>
135
+
136
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
137
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
138
+ | 1 | 63.5 | 900 | 0.7947 | 0.8240 | 0.7741 | 0.9646 | 0.1030 | 0.6047 |
139
+ | 2 | 66.2 | 500 | 0.1058 | 0.1158 | 0.0997 | 0.9357 | 0.0000 | 0.3038 |
140
+ | 3 | 65.3 | 400 | 0.8078 | 0.8419 | 0.7850 | 0.9506 | 0.0793 | 0.4778 |
141
+ | 4 | 80.7 | 650 | 0.8712 | 0.8968 | 0.8537 | 0.9638 | 0.2065 | 0.6006 |
142
+ | 5 | 108.7 | 200 | 0.9694 | 0.9824 | 0.9635 | 0.9778 | 0.5290 | 0.7625 |
143
+ | 6 | 108.3 | 100 | 0.9235 | 0.9435 | 0.9133 | 0.9665 | 0.2870 | 0.6640 |
144
+ | 7 | 178.6 | 250 | 0.8267 | 0.8514 | 0.8046 | 0.8572 | 0.3076 | 0.3760 |
145
+ | 8 | 217.2 | 150 | 0.9177 | 0.9372 | 0.9043 | 0.8987 | 0.3347 | 0.4493 |
146
+ | 9 | 144.5 | 250 | 0.8684 | 0.9066 | 0.8286 | 0.7812 | 0.0184 | 0.0160 |
147
+ | 10 | 90.2 | 150 | 0.9417 | 0.9542 | 0.9355 | 0.9800 | 0.5987 | 0.8020 |
148
+ | 11 | 76.6 | 200 | 0.9243 | 0.9453 | 0.9088 | 0.9610 | 0.3530 | 0.6250 |
149
+ | 12 | 86.0 | 150 | 0.9362 | 0.9558 | 0.9187 | 0.9440 | 0.3640 | 0.5140 |
150
+ | 13 | 58.1 | 150 | 0.8567 | 0.8872 | 0.8393 | 0.9662 | 0.1493 | 0.6033 |
151
+ | 14 | 71.9 | 150 | 0.2485 | 0.2687 | 0.2363 | 0.9418 | 0.0000 | 0.3580 |
152
+ | 15 | 68.6 | 150 | 0.8617 | 0.8970 | 0.8390 | 0.9553 | 0.0887 | 0.5220 |
153
+ | 16 | 69.2 | 200 | 0.0229 | 0.0271 | 0.0217 | 0.9280 | 0.0000 | 0.2115 |
154
+ | 17 | 112.8 | 150 | 0.1552 | 0.1765 | 0.1473 | 0.8477 | 0.0013 | 0.0133 |
155
+ | 18 | 111.0 | 150 | 0.9515 | 0.9695 | 0.9283 | 0.9027 | 0.3733 | 0.4573 |
156
+ | 19 | 89.1 | 200 | 0.0000 | 0.0000 | 0.0000 | 0.8629 | 0.0000 | 0.0005 |
157
+ | 20 | 98.9 | 250 | 0.4074 | 0.4462 | 0.3760 | 0.8706 | 0.0000 | 0.0324 |
158
+ | 21 | 92.3 | 200 | 0.8999 | 0.9182 | 0.8811 | 0.9386 | 0.4730 | 0.6105 |
159
+ | 22 | 92.3 | 200 | 0.9441 | 0.9588 | 0.9289 | 0.9469 | 0.5480 | 0.6815 |
160
+ | 23 | 142.6 | 250 | 0.9696 | 0.9815 | 0.9635 | 0.9632 | 0.5932 | 0.7012 |
161
+ | 24 | 84.7 | 200 | 0.9069 | 0.9315 | 0.8935 | 0.9665 | 0.3010 | 0.6300 |
162
+ | 25 | 108.8 | 100 | 0.9378 | 0.9470 | 0.9335 | 0.9770 | 0.7100 | 0.8230 |
163
+ | 26 | 108.8 | 100 | 0.9030 | 0.9297 | 0.8672 | 0.8755 | 0.0720 | 0.1180 |
164
+ | 27 | 152.5 | 100 | 0.2220 | 0.2417 | 0.2095 | 0.7245 | 0.0000 | 0.0030 |
165
+ | 28 | 124.8 | 200 | 0.2886 | 0.3205 | 0.2627 | 0.8064 | 0.0000 | 0.0015 |
166
+ | 29 | 134.2 | 100 | 0.8878 | 0.9270 | 0.8812 | 0.8858 | 0.2430 | 0.3030 |
167
+ | 30 | 88.8 | 150 | 0.0967 | 0.1075 | 0.0882 | 0.8798 | 0.0000 | 0.0293 |
168
+ | **mean** | | | **0.6816** | **0.7030** | **0.6662** | **0.9140** | **0.2245** | **0.4098** |
169
+
170
+ </details>
171
+
172
+ <details>
173
+ <summary><b>TUD-L</b> (3 objects, test)</summary>
174
+
175
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
176
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
177
+ | 1 | 430.3 | 200 | 0.9711 | 0.9835 | 0.9545 | 0.9439 | 0.3925 | 0.5605 |
178
+ | 2 | 175.7 | 200 | 0.9324 | 0.9547 | 0.9096 | 0.9689 | 0.2365 | 0.6570 |
179
+ | 3 | 352.4 | 200 | 0.9870 | 0.9985 | 0.9651 | 0.9633 | 0.4105 | 0.5940 |
180
+ | **mean** | | | **0.9635** | **0.9789** | **0.9431** | **0.9587** | **0.3465** | **0.6038** |
181
+
182
+ </details>
183
+
184
+ <details>
185
+ <summary><b>YCB-V</b> (21 objects, test)</summary>
186
+
187
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
188
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
189
+ | 1 | 172.1 | 300 | 0.9468 | 0.9800 | 0.9470 | 0.9587 | 0.1273 | 0.5660 |
190
+ | 2 | 269.6 | 225 | 0.9967 | 0.9996 | 0.9943 | 0.9876 | 0.6609 | 0.7849 |
191
+ | 3 | 198.4 | 375 | 0.9925 | 0.9975 | 0.9906 | 0.9956 | 0.6568 | 0.8539 |
192
+ | 4 | 120.5 | 450 | 0.9518 | 0.9670 | 0.9436 | 0.9778 | 0.4098 | 0.7502 |
193
+ | 5 | 196.5 | 150 | 0.9512 | 0.9622 | 0.9373 | 0.9408 | 0.3940 | 0.5693 |
194
+ | 6 | 89.8 | 300 | 0.7816 | 0.8251 | 0.7290 | 0.8697 | 0.0230 | 0.1147 |
195
+ | 7 | 142.5 | 75 | 0.9780 | 0.9923 | 0.9743 | 0.9943 | 0.3600 | 0.8467 |
196
+ | 8 | 114.1 | 75 | 0.9910 | 1.0000 | 0.9863 | 1.0000 | 0.5920 | 0.9480 |
197
+ | 9 | 129.5 | 225 | 0.8569 | 0.8846 | 0.8296 | 0.9054 | 0.2316 | 0.4089 |
198
+ | 10 | 197.8 | 150 | 0.7247 | 0.7437 | 0.6943 | 0.9063 | 0.0400 | 0.2027 |
199
+ | 11 | 259.5 | 225 | 0.7109 | 0.7407 | 0.6770 | 0.6787 | 0.0262 | 0.0236 |
200
+ | 12 | 259.6 | 300 | 0.7168 | 0.7312 | 0.6992 | 0.8213 | 0.2130 | 0.2900 |
201
+ | 13 | 161.9 | 150 | 0.8455 | 0.8883 | 0.8307 | 0.8267 | 0.0587 | 0.0527 |
202
+ | 14 | 125.0 | 150 | 0.9562 | 0.9832 | 0.9252 | 0.9317 | 0.0500 | 0.3567 |
203
+ | 15 | 226.2 | 300 | 0.9940 | 0.9972 | 0.9878 | 0.9849 | 0.5987 | 0.7610 |
204
+ | 16 | 237.3 | 75 | 0.6860 | 0.7153 | 0.6660 | 0.8560 | 0.0000 | 0.0320 |
205
+ | 17 | 204.0 | 75 | 0.9407 | 0.9567 | 0.9337 | 0.9587 | 0.1373 | 0.5240 |
206
+ | 18 | 121.4 | 150 | 0.9830 | 0.9893 | 0.9788 | 0.9983 | 0.5080 | 0.9033 |
207
+ | 19 | 174.7 | 150 | 0.9033 | 0.9292 | 0.8903 | 0.9232 | 0.0753 | 0.4100 |
208
+ | 20 | 217.1 | 150 | 0.9152 | 0.9480 | 0.8945 | 0.9010 | 0.1467 | 0.2180 |
209
+ | 21 | 102.9 | 75 | 0.3873 | 0.4257 | 0.3460 | 0.8553 | 0.0000 | 0.0093 |
210
+ | **mean** | | | **0.8671** | **0.8884** | **0.8503** | **0.9177** | **0.2528** | **0.4584** |
211
+
212
+ </details>
213
+
214
+ <details>
215
+ <summary><b>HB</b> (33 objects, val_primesense)</summary>
216
+
217
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC |
218
+ |----|-----------|------|---------|-----------|----------|
219
+ | 1 | 232.6 | 680 | 0.9906 | 0.9962 | 0.9803 |
220
+ | 2 | 257.4 | 340 | 1.0000 | 1.0000 | 1.0000 |
221
+ | 3 | 166.5 | 1020 | 0.9737 | 0.9834 | 0.9640 |
222
+ | 4 | 179.0 | 680 | 0.9849 | 0.9908 | 0.9793 |
223
+ | 5 | 205.4 | 680 | 0.9864 | 0.9899 | 0.9825 |
224
+ | 6 | 121.4 | 340 | 0.9741 | 0.9968 | 0.9471 |
225
+ | 7 | 263.7 | 340 | 0.9984 | 1.0000 | 0.9940 |
226
+ | 8 | 186.8 | 680 | 0.9931 | 0.9968 | 0.9903 |
227
+ | 9 | 166.6 | 680 | 0.9139 | 0.9348 | 0.8892 |
228
+ | 10 | 180.8 | 680 | 0.9338 | 0.9546 | 0.9307 |
229
+ | 11 | 238.5 | 340 | 0.9153 | 0.9433 | 0.9260 |
230
+ | 12 | 156.9 | 1360 | 0.9688 | 0.9823 | 0.9537 |
231
+ | 13 | 145.3 | 1020 | 0.9582 | 0.9733 | 0.9455 |
232
+ | 14 | 243.7 | 680 | 0.9922 | 0.9965 | 0.9857 |
233
+ | 15 | 113.0 | 1700 | 0.9483 | 0.9611 | 0.9374 |
234
+ | 16 | 101.6 | 1020 | 0.9519 | 0.9679 | 0.9296 |
235
+ | 17 | 132.8 | 1360 | 0.9448 | 0.9571 | 0.9293 |
236
+ | 18 | 211.1 | 680 | 0.1994 | 0.2258 | 0.1693 |
237
+ | 19 | 185.6 | 680 | 0.9987 | 0.9993 | 0.9982 |
238
+ | 20 | 244.8 | 340 | 1.0000 | 1.0000 | 1.0000 |
239
+ | 21 | 212.6 | 340 | 0.9979 | 0.9993 | 0.9963 |
240
+ | 22 | 190.2 | 1360 | 0.9874 | 0.9933 | 0.9823 |
241
+ | 23 | 233.9 | 1020 | 0.9987 | 0.9997 | 0.9971 |
242
+ | 24 | 252.3 | 340 | 0.9940 | 0.9999 | 0.9889 |
243
+ | 25 | 202.9 | 680 | 0.9753 | 0.9853 | 0.9612 |
244
+ | 26 | 183.8 | 680 | 0.9785 | 0.9893 | 0.9613 |
245
+ | 27 | 264.4 | 340 | 0.9989 | 0.9998 | 0.9983 |
246
+ | 28 | 477.5 | 340 | 1.0000 | 1.0000 | 1.0000 |
247
+ | 29 | 198.0 | 680 | 0.9720 | 0.9970 | 0.9489 |
248
+ | 30 | 416.2 | 340 | 1.0000 | 1.0000 | 1.0000 |
249
+ | 31 | 158.0 | 340 | 0.9912 | 0.9971 | 0.9817 |
250
+ | 32 | 201.8 | 680 | 0.9610 | 0.9701 | 0.9487 |
251
+ | 33 | 187.2 | 680 | 0.9945 | 0.9974 | 0.9916 |
252
+ | **mean** | | | **0.9538** | **0.9630** | **0.9451** |
253
+
254
+ </details>
255
+
256
+ <details>
257
+ <summary><b>ITODD</b> (28 objects, val)</summary>
258
+
259
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
260
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
261
+ | 1 | 64.1 | 4 | 0.9125 | 0.9437 | 0.8562 | 0.5500 | 0.0000 | 0.0000 |
262
+ | 2 | 51.5 | 3 | 0.4833 | 0.5000 | 0.4500 | 0.7167 | 0.0000 | 0.0000 |
263
+ | 3 | 142.2 | 4 | 0.0000 | 0.0000 | 0.0000 | 0.2750 | 0.0000 | 0.0000 |
264
+ | 4 | 139.4 | 3 | 0.2583 | 0.3000 | 0.2250 | 0.3083 | 0.0000 | 0.0000 |
265
+ | 5 | 158.6 | 6 | 0.7167 | 0.7583 | 0.6833 | 0.2500 | 0.0000 | 0.0000 |
266
+ | 6 | 85.3 | 5 | 0.9650 | 0.9750 | 0.9650 | 0.9750 | 0.1800 | 0.6000 |
267
+ | 7 | 38.5 | 5 | 0.0000 | 0.0200 | 0.0000 | 0.8250 | 0.0000 | 0.0000 |
268
+ | 8 | 68.9 | 3 | 0.8417 | 0.8917 | 0.7917 | 0.5083 | 0.0000 | 0.0000 |
269
+ | 9 | 94.8 | 3 | 0.0000 | 0.0000 | 0.0000 | 0.4000 | 0.0000 | 0.0000 |
270
+ | 10 | 55.7 | 4 | 0.9563 | 0.9688 | 0.9313 | 0.8938 | 0.0000 | 0.1500 |
271
+ | 11 | 140.1 | 5 | 0.5950 | 0.6300 | 0.5850 | 0.4550 | 0.0600 | 0.0000 |
272
+ | 12 | 107.7 | 4 | 0.0000 | 0.0000 | 0.0000 | 0.1437 | 0.0000 | 0.0000 |
273
+ | 13 | 128.1 | 4 | 0.9750 | 0.9875 | 0.9563 | 0.7438 | 0.1750 | 0.1000 |
274
+ | 14 | 102.9 | 3 | 0.0000 | 0.0000 | 0.0000 | 0.3917 | 0.0000 | 0.0000 |
275
+ | 15 | 114.2 | 3 | 0.9583 | 0.9833 | 0.9167 | 0.4750 | 0.0333 | 0.0000 |
276
+ | 16 | 193.1 | 3 | 0.9833 | 1.0000 | 0.9667 | 0.5917 | 0.2333 | 0.0000 |
277
+ | 17 | 77.8 | 3 | 0.7333 | 0.7833 | 0.6833 | 0.6167 | 0.0000 | 0.0000 |
278
+ | 18 | 108.5 | 3 | 0.2083 | 0.2417 | 0.1917 | 0.6583 | 0.0000 | 0.0000 |
279
+ | 19 | 121.4 | 3 | 0.7917 | 0.8417 | 0.7500 | 0.5083 | 0.0000 | 0.0000 |
280
+ | 20 | 122.0 | 4 | 0.9812 | 0.9875 | 0.9750 | 0.9062 | 0.6250 | 0.4250 |
281
+ | 21 | 171.2 | 3 | 0.9917 | 1.0000 | 0.9833 | 0.7833 | 0.5333 | 0.1333 |
282
+ | 22 | 267.5 | 3 | 0.9333 | 0.9667 | 0.8833 | 0.3333 | 0.0000 | 0.0000 |
283
+ | 23 | 56.9 | 1 | 1.0000 | 1.0000 | 0.9750 | 0.9750 | 0.4000 | 0.7000 |
284
+ | 24 | 65.0 | 6 | 0.3333 | 0.3583 | 0.3292 | 0.8708 | 0.0000 | 0.0000 |
285
+ | 25 | 48.5 | 6 | 0.9375 | 0.9625 | 0.9083 | 0.7917 | 0.0000 | 0.3000 |
286
+ | 26 | 66.8 | 6 | 0.9792 | 0.9875 | 0.9667 | 0.9458 | 0.3500 | 0.4333 |
287
+ | 27 | 55.7 | 5 | 0.7200 | 0.7750 | 0.6700 | 0.6800 | 0.0000 | 0.0000 |
288
+ | 28 | 24.1 | 18 | 0.5306 | 0.5681 | 0.5014 | 0.8917 | 0.0000 | 0.0611 |
289
+ | **mean** | | | **0.6352** | **0.6582** | **0.6123** | **0.6237** | **0.0925** | **0.1037** |
290
+
291
+ </details>
292
+
293
+ <details>
294
+ <summary><b>IC-BIN</b> (2 objects, test)</summary>
295
+
296
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
297
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
298
+ | 1 | 136.6 | 1800 | 0.8560 | 0.8871 | 0.8380 | 0.9422 | 0.2424 | 0.4592 |
299
+ | 2 | 220.6 | 450 | 0.8723 | 0.8971 | 0.8364 | 0.8962 | 0.0671 | 0.2191 |
300
+ | **mean** | | | **0.8641** | **0.8921** | **0.8372** | **0.9192** | **0.1548** | **0.3391** |
301
+
302
+ </details>
303
+
304
+ <details>
305
+ <summary><b>HOPE</b> (28 objects, val)</summary>
306
+
307
+ | ID | Diam (mm) | N GT | ADD-AUC | ADD-S-AUC | MSSD-AUC | MSPD-AUC | AR<sub>MSSD</sub> | AR<sub>MSPD</sub> |
308
+ |----|-----------|------|---------|-----------|----------|----------|-----------|-----------|
309
+ | 1 | 107.9 | 45 | 0.8717 | 0.8867 | 0.8644 | 0.9417 | 0.4178 | 0.5844 |
310
+ | 2 | 153.9 | 45 | 0.8994 | 0.9150 | 0.8733 | 0.8478 | 0.3133 | 0.4000 |
311
+ | 3 | 115.0 | 45 | 0.8622 | 0.8778 | 0.8533 | 0.9439 | 0.4378 | 0.6200 |
312
+ | 4 | 89.6 | 20 | 0.9525 | 0.9713 | 0.9400 | 0.9475 | 0.3050 | 0.5300 |
313
+ | 5 | 98.0 | 30 | 0.8350 | 0.8458 | 0.8250 | 0.9417 | 0.5300 | 0.6633 |
314
+ | 6 | 208.7 | 25 | 0.9910 | 0.9960 | 0.9880 | 0.9370 | 0.5920 | 0.4880 |
315
+ | 7 | 89.6 | 40 | 0.8325 | 0.8663 | 0.8150 | 0.9187 | 0.1250 | 0.3600 |
316
+ | 8 | 115.6 | 25 | 0.8990 | 0.9190 | 0.8780 | 0.8470 | 0.2760 | 0.3600 |
317
+ | 9 | 206.0 | 25 | 0.8610 | 0.8910 | 0.8210 | 0.7250 | 0.2000 | 0.1880 |
318
+ | 10 | 89.7 | 35 | 0.6257 | 0.6493 | 0.5979 | 0.8479 | 0.1571 | 0.2286 |
319
+ | 11 | 153.9 | 40 | 0.9238 | 0.9338 | 0.9094 | 0.9256 | 0.5225 | 0.6525 |
320
+ | 12 | 207.1 | 50 | 0.9535 | 0.9710 | 0.9310 | 0.7805 | 0.3500 | 0.2800 |
321
+ | 13 | 153.4 | 30 | 0.9767 | 0.9917 | 0.9550 | 0.8525 | 0.3767 | 0.3733 |
322
+ | 14 | 204.5 | 25 | 0.9660 | 0.9760 | 0.9500 | 0.8920 | 0.4120 | 0.3520 |
323
+ | 15 | 75.7 | 35 | 0.8814 | 0.9150 | 0.8521 | 0.8929 | 0.0657 | 0.3943 |
324
+ | 16 | 161.5 | 50 | 0.9560 | 0.9660 | 0.9420 | 0.8980 | 0.5120 | 0.5220 |
325
+ | 17 | 205.7 | 25 | 0.7150 | 0.7290 | 0.6790 | 0.7060 | 0.1960 | 0.1840 |
326
+ | 18 | 122.8 | 30 | 0.9658 | 0.9783 | 0.9458 | 0.9133 | 0.4167 | 0.4467 |
327
+ | 19 | 89.2 | 20 | 0.9650 | 0.9800 | 0.9550 | 0.9750 | 0.3600 | 0.7350 |
328
+ | 20 | 89.9 | 30 | 0.7575 | 0.7900 | 0.7350 | 0.9050 | 0.0900 | 0.3400 |
329
+ | 21 | 89.2 | 55 | 0.8286 | 0.8509 | 0.8141 | 0.9223 | 0.2855 | 0.5655 |
330
+ | 22 | 152.4 | 20 | 0.7838 | 0.8137 | 0.7475 | 0.8462 | 0.0650 | 0.1350 |
331
+ | 23 | 151.3 | 40 | 0.8206 | 0.8475 | 0.7900 | 0.7162 | 0.0975 | 0.1500 |
332
+ | 24 | 151.3 | 10 | 1.0000 | 1.0000 | 0.9975 | 0.9750 | 0.8000 | 0.6500 |
333
+ | 25 | 252.8 | 35 | 0.9786 | 0.9836 | 0.9714 | 0.9257 | 0.5286 | 0.5400 |
334
+ | 26 | 107.1 | 35 | 0.7529 | 0.7679 | 0.7457 | 0.9364 | 0.3771 | 0.5800 |
335
+ | 27 | 76.1 | 35 | 0.7736 | 0.7971 | 0.7636 | 0.9443 | 0.1629 | 0.4886 |
336
+ | 28 | 82.9 | 20 | 0.8962 | 0.9150 | 0.8875 | 0.9525 | 0.3550 | 0.5750 |
337
+ | **mean** | | | **0.8759** | **0.8937** | **0.8581** | **0.8878** | **0.3331** | **0.4424** |
338
+
339
+ </details>
340
 
341
  ## File Structure
342
 
343
  ```
344
  {dataset}/obj_{NNNNNN}/
345
+ best_coco_AP_epoch_NNN.pth # DOPER-t keypoint checkpoint
346
  keypoints_3d.json # 17 symmetry-aware 3D keypoints (mm)
347
  bop_summary.json # Evaluation metrics
348
+ vis_grid.jpg # Qualitative results (GT bbox eval)
349
  ```
350
 
351
+ Detector checkpoints and BOP submission CSVs are in the `bop_submission/` directory.
352
+
353
  ## Usage
354
 
355
  ```python
356
  from huggingface_hub import hf_hub_download
357
 
358
+ # Download keypoint model + 3D keypoints for ycbv object 1
359
+ ckpt = hf_hub_download("TontonTremblay/DOPER_BOP", "ycbv/obj_000001/best_coco_AP_epoch_200.pth", repo_type="dataset")
360
  kpts = hf_hub_download("TontonTremblay/DOPER_BOP", "ycbv/obj_000001/keypoints_3d.json", repo_type="dataset")
361
  ```
362
+
363
+ ## BOP Submission
364
+
365
+ Pre-computed detection+pose results for all 9 datasets in BOP CSV format:
366
+
367
+ - `bop_submission/doper-t_bop_results_det_v2.zip` -- RTMDet + DOPER-t + PnP (no GT bbox)
368
+
369
+ ## Citation
370
+
371
+ If you use these models, please cite the DOPER project and the [BOP benchmark](https://bop.felk.cvut.cz/).