File size: 2,246 Bytes
5e656e6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: other
tags:
  - keypoint-matching
  - superpoint
  - superglue
  - inverse-graphics
---

# Weights for *Structure from Duplicates* (SfD)

Pretrained checkpoints used by the SfD preprocessing pipeline
([Tianhang-Cheng/SfD](https://github.com/Tianhang-Cheng/SfD)). They are hosted here instead of in the
git repository to keep clones small.

| file | size | what it is |
| --- | --- | --- |
| `keypoint_matching/superpoint_v1.pth` | 5 MB | SuperPoint detector/descriptor |
| `keypoint_matching/superglue_indoor.pth` | 46 MB | SuperGlue matcher, indoor weights |
| `keypoint_matching/superglue_outdoor.pth` | 46 MB | SuperGlue matcher, outdoor weights |

## Usage

From a clone of the repo, all three land in `preprocess/keypoint_matching/weights/`:

```bash
python download_assets.py --weights
```

The pipeline also fetches them lazily the first time the matcher is constructed, so this step is
optional.

Directly:

```python
from huggingface_hub import hf_hub_download

hf_hub_download("TianhangCheng7/DuplicateWeight", "keypoint_matching/superpoint_v1.pth")
```

## License / attribution

SuperPoint and SuperGlue were released by Magic Leap, Inc. and are mirrored here under the original
[SuperGluePretrainedNetwork](https://github.com/magicleap/SuperGluePretrainedNetwork) terms, which
allow **non-commercial research use only**. Please cite the original works:

```
@inproceedings{detone2018superpoint,
  title={SuperPoint: Self-Supervised Interest Point Detection and Description},
  author={DeTone, Daniel and Malisiewicz, Tomasz and Rabinovich, Andrew},
  booktitle={CVPR Deep Learning for Visual SLAM Workshop},
  year={2018}
}
@inproceedings{sarlin2020superglue,
  title={SuperGlue: Learning Feature Matching with Graph Neural Networks},
  author={Sarlin, Paul-Edouard and DeTone, Daniel and Malisiewicz, Tomasz and Rabinovich, Andrew},
  booktitle={CVPR},
  year={2020}
}
```

If you use SfD:

```
@inproceedings{cheng2023structure,
  title={Structure from Duplicates: Neural Inverse Graphics from a Pile of Objects},
  author={Cheng, Tianhang and Ma, Wei-Chiu and Guan, Kaiyu and Torralba, Antonio and Wang, Shenlong},
  booktitle={Thirty-seventh Conference on Neural Information Processing Systems},
  year={2023}
}
```