Base library for visuo-tactile shape mapping with gelsight and depth camera. For more information, consider our work Efficient shape mapping through dense touch and vision.
The library is still being actively updated, please check back soon for a stable version!
- Initial code upload
- Data upload
- Compilation fixes
- Enhanced documentation
matlab/
: Executable and utility scripts for mappingshape-map/
: 3-D and 2-D mapping scripts for both simulated and real-world datautils/
: helpers scripts
gpfactor/
: cpp header wrapped with the gtsam + MATLAB library
- GTSAM: Checkout gtsam 4.0.2
master
branch (notdevelop
):- Using the default CMakeLists without modification
cmake -DGTSAM_USE_SYSTEM_EIGEN=ON -DGTSAM_WITH_TBB=OFF -DGTSAM_INSTALL_MATLAB_TOOLBOX=ON .. sudo make install -j
- MATLAB toolbox check
- Open MATLAB
- add
addpath /usr/local/gtsam_toolbox
to thestartup.m
- Test
gtsamExamples
- Using the default CMakeLists without modification
- C++ header compilation:
mkdir build cd build cmake -DGPFACTOR_BUILD_TOOLBOX:OPTION=ON -DCMAKE_BUILD_TYPE=Release -DGTSAM_TOOLBOX_INSTALL_PATH:PATH=/usr/local/gtsam_toolbox .. sudo make install
sudo ldconfig
in terminal- open MATLAB here
- Run
.m
files
Feel free to use the library as you please. If you find it helpful, please consider referencing:
@misc{suresh2021efficient,
title={Efficient shape mapping through dense touch and vision},
author={Sudharshan Suresh and Zilin Si and Joshua G. Mangelson and Wenzhen Yuan and Michael Kaess},
year={2021},
eprint={2109.09884},
archivePrefix={arXiv},
primaryClass={cs.RO}
}