chrischoy commited on
Commit
a373c87
·
verified ·
1 Parent(s): 4068a14

Add model card

Browse files
Files changed (1) hide show
  1. README.md +60 -0
README.md ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ tags:
4
+ - point-cloud
5
+ - 3d-registration
6
+ - geometric-features
7
+ - pytorch
8
+ - minkowski-engine
9
+ ---
10
+
11
+ # Fully Convolutional Geometric Features (FCGF)
12
+
13
+ Pretrained weights for **Fully Convolutional Geometric Features**, ICCV 2019.
14
+
15
+ - Paper: [Fully Convolutional Geometric Features](https://node1.chrischoy.org/data/publications/fcgf/fcgf.pdf)
16
+ - Code: [chrischoy/FCGF](https://github.com/chrischoy/FCGF)
17
+
18
+ FCGF extracts dense 3D geometric features in a single forward pass through a 3D fully-convolutional network, built on [MinkowskiEngine](https://github.com/NVIDIA/MinkowskiEngine). These checkpoints are the official weights released with the paper, previously hosted on `node1.chrischoy.org` and mirrored here.
19
+
20
+ ## Model Zoo
21
+
22
+ All models use the `ResUNetBN2C` backbone.
23
+
24
+ | File | Normalized Feature | Dataset | Voxel Size | Feature Dimension | Performance |
25
+ |:-----------------------------------------------|:-------------------:|:-------:|:-------------:|:-----------------:|:--------------------------:|
26
+ | [2019-08-19_06-17-41.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/2019-08-19_06-17-41.pth) | True | 3DMatch | 2.5cm (0.025) | 32 | FMR: 0.9578 +- 0.0272 |
27
+ | [2019-09-18_14-15-59.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/2019-09-18_14-15-59.pth) | True | 3DMatch | 2.5cm (0.025) | 16 | FMR: 0.9442 +- 0.0345 |
28
+ | [2019-08-16_19-21-47.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/2019-08-16_19-21-47.pth) | True | 3DMatch | 5cm (0.05) | 32 | FMR: 0.9372 +- 0.0332 |
29
+ | [2019-07-31_19-30-19.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/2019-07-31_19-30-19.pth) | False | KITTI | 20cm (0.2) | 32 | RTE: 0.0534m, RRE: 0.1704° |
30
+ | [2019-07-31_19-37-00.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/2019-07-31_19-37-00.pth) | False | KITTI | 30cm (0.3) | 32 | RTE: 0.0607m, RRE: 0.2280° |
31
+ | [KITTI-v0.3-ResUNetBN2C-conv1-5-nout16.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/KITTI-v0.3-ResUNetBN2C-conv1-5-nout16.pth) | True | KITTI | 30cm (0.3) | 16 | RTE: 0.0670m, RRE: 0.2295° |
32
+ | [KITTI-v0.3-ResUNetBN2C-conv1-5-nout32.pth](https://huggingface.co/chrischoy/FCGF/resolve/main/KITTI-v0.3-ResUNetBN2C-conv1-5-nout32.pth) | True | KITTI | 30cm (0.3) | 32 | RTE: 0.0639m, RRE: 0.2253° |
33
+
34
+ `redkitchen-20.ply` is the sample point cloud used by [`demo.py`](https://github.com/chrischoy/FCGF/blob/master/demo.py) in the FCGF repo.
35
+
36
+ ## Usage
37
+
38
+ ```python
39
+ import torch
40
+ from urllib.request import urlretrieve
41
+
42
+ urlretrieve(
43
+ "https://huggingface.co/chrischoy/FCGF/resolve/main/2019-09-18_14-15-59.pth",
44
+ "ResUNetBN2C-16feat-3conv.pth")
45
+
46
+ checkpoint = torch.load("ResUNetBN2C-16feat-3conv.pth")
47
+ ```
48
+
49
+ See [chrischoy/FCGF](https://github.com/chrischoy/FCGF) for the full model definition (`ResUNetBN2C`) and feature extraction code.
50
+
51
+ ## Citation
52
+
53
+ ```bibtex
54
+ @inproceedings{FCGF2019,
55
+ author = {Christopher Choy and Jaesik Park and Vladlen Koltun},
56
+ title = {Fully Convolutional Geometric Features},
57
+ booktitle = {ICCV},
58
+ year = {2019},
59
+ }
60
+ ```