| 1 |
LINEMOD |
 |
2012 |
Instance |
6D pose + Depth for one object |
manual |
15 video with 18,273 frames |
R |
15 |
Moving cam |
VP + C + TL |
Yes |
Limited |
Cluttered |
Household |
- |
| 2 |
Occluded-LINEMOD |
 |
2014 |
Instance |
6D Pose + Depth |
semi-auto |
8 models in 1,214 frames |
R |
8 |
Moving cam |
VP + C + TL + O |
NO |
Limited |
Cluttered |
Household |
- |
| 3 |
APC |
 |
2016 |
Instance |
RGB-D + mask |
semi-manual |
10,000 RGB-D images |
R |
24 |
Moving cam |
VP + C |
- |
- |
Cluttered |
Household |
- |
| 4 |
T-LESS |
 |
2017 |
Instance |
6D Pose + Depth + 3D model |
semi-auto |
20 videos with ~49K frames |
R |
30 |
Moving cam |
VP + C + TL + O + MI + SLD |
- |
High |
Cluttered |
Industrial |
- |
| 5 |
YCB-Video |
 |
2018 |
Instance |
6D Pose + Depth + Mask |
semi-auto |
92 videos with 133,827 frames |
R |
21 |
Moving cam |
VP+ O + image noise |
Yes |
Limited |
Cluttered |
Household |
- |
| 6 |
Falling Things (FAT) |
 |
2018 |
Instance |
3D poses, mask, and 2D/3D bounding box coordinates for all objects |
auto |
60.000 images |
S |
21 |
Stereo (3 scenes and 5 locations within each scene) |
VP + C |
Yes |
- |
Cluttered |
Household |
- |
| 7 |
Fraunhofer IPA (BIN-P) |
 |
2019 |
Instance |
RGB-D + 3D point cloud + mask + Pose |
semi-auto |
520 real and 206.000 synth images |
R&S |
8+2 |
- |
VP + SC + SO + MI + BP |
- |
High |
Stacked |
Industrial |
- |
| 8 |
NOCS |
 |
2019 |
Category |
6D pose + Depth + NOCS |
semi-auto |
300K composited images and 8K real images |
R&S |
(6) |
31 Scenes |
VP + O + C + Category challenges |
- |
- |
- |
Household |
- |
| 9 |
CAMERA |
 |
2019 |
Category |
RGB-D + 6D pose |
semi-auto |
300.000 frames |
R&S |
42(6) |
31 Scenes |
VP + O + C + Category challenges |
Yes |
Limited |
Cluttered |
Household |
- |
| 10 |
ObjectSynth |
 |
2019 |
Instance |
2D bounding boxes, masks and 6D poses of the visible object instances |
automated |
Synthetic images with 600,000 frames |
S |
39 |
6 Scenes & 200 views |
VP + O + C |
Yes |
- |
- |
Household |
- |
| 11 |
HomebrewedDB |
 |
2019 |
Instance |
RGB-D images, surface normal vectors , 3D model |
auto |
17420 images (1340/scene) |
S |
33 |
13 Scenes |
VP + O + C + Light |
Yes |
- |
Cluttered |
Household and Industrial |
- |
| 12 |
GraspNet-1B |
 |
2020 |
Instance |
6D pose + Depth + Mask |
semi-auto |
190 videos with 97,280 frames |
R |
88 |
190 Scenes |
VP + O + C |
- |
- |
Cluttered |
Household |
- |
| 13 |
RobotP |
 |
2020 |
Instance |
RGB-D images, 6D pose, 3D point clouds, 3D models and object masks and bounding boxes |
semi-auto |
4000 images (500/object) |
S |
8 |
Moving cam (sim-robot) |
VP + TL + Light + O |
Yes |
- |
Cluttered |
Household |
- |
| 14 |
HOPE |
 |
2021 |
Instance |
6D pose + RGB-D |
manual |
RGBD images and video sequences (2038 frames) |
R |
28 |
50 scenes from 10 household/office environments |
VP + C + O + MI + Light |
Yes |
- |
Cluttered |
Household |
- |
| 15 |
MetaGraspNet |
 |
2021 |
Instance |
Depth + Mask |
automated |
100,000 RGB-D |
S |
25 |
11,000 scenes |
VP + O + C + MI |
- |
- |
Stacked |
Household |
- |
| 16 |
SynPick |
 |
2021 |
Instance |
RGB + mask + object contours with grasp point + graph scene layout |
automated |
Synthetic videos with 503232 frames |
S |
21 |
300 Scenes (dynamic) & 3 views (multi) |
Vp + O |
Yes |
- |
- |
Household |
- |
| 17 |
StereOBJ-1M D |
-- |
2021 |
- |
- |
- |
396K frames |
R |
18 |
183 scenes constructed in 11 different environments |
-- |
- |
- |
- |
- |
- |
| 18 |
DoPose |
 |
2022 |
Instance |
Depth, 6D Pose, mask and 3D model of all objects |
semi-auto |
3325 images |
R |
19 |
301 Scenes and 3325 view images |
VP + O + C |
- |
- |
Cluttered |
Household |
- |
| 19 |
(Ours) |
-- |
2022 |
I & C |
6D Pose, mask, 2D&3D Bbox, RGB-D, 3D point cloud, 3D model, Normal map, reflectance map, indintifier map infra1 &2, (RGB based on Depth image) |
automated |
?? |
S |
24(15) & (24) |
(X) views |
Almost all challenges (-TL) |
Yes |
High |
Cluttered & (~Stacked) |
Household (Fruits & Vegetables) |
Mass, Color, Texture, Size, Lights, Camera poses |
