maolingao / SIGMA

the official implementation of the paper "SIGMA: Scale-Invariant Global Sparse Shape Matching" (ICCV 2023)

SIGMA: Scale-Invariant Global Sparse Shape Matching

Paper-CVF | Paper-arXiv |

Citation

If you find our work useful in your research, please consider citing:

@inproceedings{gao2023sigma,
    author 	= { Maolin Gao and Paul Roetzer and Marvin Eisenberger and Zorah L\"ahner and Michael Moeller and Daniel Cremers and Florian Bernard },
    title 	= { {SIGMA}: Scale-Invariant Global Sparse Shape Matching},
    booktitle = {International Conference on Computer Vision (ICCV)},
    year 	= 2023,
    keywords = {Shape Analysis, Geometry Processing, Global Optimisation, Mixed-Integer Programming, Laplace-Betrami}
}

Note: The initial public release in this repository corresponds to the code version evluated in the ICCV'23 paper, after refactoring and cleanup. As the code evolves, runtime differences might become larger.

Running SIGMA

Please install YALMIP and MOSEK to be able to run the optimisation. At the time when this code is developed, the following package/software version are used:

1. YALMIP Version R202110331
2. MOSEK Version 10.0.35 (Build date: 2023-2-3 09:48:18)
3. gptoolbox is inlcuded under code/external/gptoolbox-master/ for your convenience, however you might want to use a more up-to-date version.

Then please refer to demo1.m for an example of TOSCA cat and demo2.m for an example of SHREC20 Non-Isometric camel to dog.

Note: The code was developed under MacOS and Ubuntu.

Acknowledgement

We thank the authors of BCICP (Jing Ren, Adrien Poulenard, Peter Wonka and Maks Ovsjanikov.), from which we took inspiration of the orientation preservation term. Their code is provided as a copy in this repository for covenience. Some of their functions are adapted for our purposes, such as visualisation functions (code/external/bcicp-repo/).

We thank the authours of gptoolbox (Alec Jacobson and co-authers) for providing the wonderfull geometry processing toolbox (code/external/gptoolbox-master/).

We also thank Dirk-Jan Kroon for open source their published function patchnormals.m.

License

The code of the IsoMush project is licensed under a BSD 3-Clause License.

Please also consider licenses of used third-party codes and libraries. See folder code/external.