ASGrasp: Generalizable Transparent Object Reconstruction and 6-DoF Grasp Detection from RGB-D Active Stereo Camera
This repository contains the source code for our paper: ASGrasp: Generalizable Transparent Object Reconstruction and 6-DoF Grasp Detection from RGB-D Active Stereo Camera
For more information, please visit our project page.
The code has been tested with Python 3.8
, PyTorch 1.9.1
and Cuda 11.1
Please refer to https://github.com/rhett-chen/graspness_implementation to install GSNet
Please download the checkpoints to 'checkpoints/'
Please run
python infer_mvs_2layer_gsnet.py
Our code is based on these wonderful repos, we appreciate their great works!
If you have any questions, please open a github issue or contact us:
Jun Shi: jun7.shi@samsung.com, Yong A: yong.a@samsung.com, He Wang: hewang@pku.edu.cn