https://github.com/eliphatfs/skeletonmerger
Code repository for paper `Skeleton Merger: an Unsupervised Aligned Keypoint Detector`.
https://github.com/eliphatfs/skeletonmerger
3d-computer-vision cvpr2021 pytorch unsupervised-learning
Last synced: 5 months ago
JSON representation
Code repository for paper `Skeleton Merger: an Unsupervised Aligned Keypoint Detector`.
- Host: GitHub
- URL: https://github.com/eliphatfs/skeletonmerger
- Owner: eliphatfs
- License: mit
- Created: 2021-03-12T04:11:45.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2022-11-20T03:24:13.000Z (almost 3 years ago)
- Last Synced: 2023-03-01T11:25:15.477Z (over 2 years ago)
- Topics: 3d-computer-vision, cvpr2021, pytorch, unsupervised-learning
- Language: Python
- Homepage:
- Size: 112 KB
- Stars: 48
- Watchers: 2
- Forks: 11
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Skeleton Merger
Skeleton Merger, an Unsupervised Aligned Keypoint Detector.
The paper is available at [https://arxiv.org/abs/2103.10814](https://arxiv.org/abs/2103.10814).
[
](_readme_resources/intropic.jpg)
**Update Aug. 6th:** The point cloud visualizer is now released! See [https://github.com/eliphatfs/PointCloudVisualizer](https://github.com/eliphatfs/PointCloudVisualizer).
## A map of the repository
+ The `merger/pointnetpp` folder contains the [Pytorch Implementation of PointNet and PointNet++](https://github.com/yanx27/Pointnet_Pointnet2_pytorch) repository with some minor changes. It is adapted to make compatible relative imports.
+ The `merger/composed_chamfer.py` file contains an efficient implementation of proposed Composite Chamfer Distance (CCD).
+ The `merger/data_flower.py` file is for data loading and preprocessing.
+ The `merger/merger_net.py` file contains the `Skeleton Merger` implementation.
+ The root folder contains several scripts for training and testing.
You can find a pre-trained model on chairs from ShapeNetCore [here](https://github.com/eliphatfs/SkeletonMerger/issues/8). Notice that axis order (e.g., gravity axis may be either y or z) and scaling may vary between datasets, so it is recommended to train a model locally from scratch if you need to use Skeleton Merger. It's fast! Skeleton Merger usually gives reasonable results within 5-10 epochs, which only takes minutes on ShapeNetCore-scale datasets with a GTX 1080. (For full power of the model you still need to train for 50-100 epochs and do some epoch selection by validation error or by the downstream task.)
## Usage of script files
Usage of the script files, together with a brief description of data format, are available through the `-h` command line option.
## Dataset
The ShapeNetCore.v2 dataset used in the paper is available from the [Point Cloud Datasets](https://github.com/AnTao97/PointCloudDatasets) repository.