Code release for "Elastic Tactile Simulation Towards Tactile-Visual Perception", in ACMMM (Oral) 2021.
By Yikai Wang, Wenbing Huang, Bin Fang, Fuchun Sun, Chang Li.
python==3.7.6
taichi==0.6.32
pytorch==1.6.0
pytorch3d==0.2.5
imageio==2.9.0
open3d==0.11.0
opencv-python==4.4.0
openexr==2.4.1
We provide two objects (torus and gear) as examples.
First, transform mesh to voxels
python mesh2voxel.py --path obj_path # e.g., obj/torus.obj
Then perform EIP, for example,
python show_torus_vertical.py --name exp_name
Will obtain fine-grained tacile patterns like,
Simulated particles of the tactile sensor will be automatically saved, which can be visualized with Meshlab,
Particles can be reconstructed to meshes by Meshlab (e.g., Surface Reconstruction). We adopt Mitsuba for rendering.
Similarly, performing the vertical contact on the gear object (show_gear_vertical.py) provides the following deformation (after rendering),
We thank a lot for the flexible codebase of Taichi.
If you find our work useful for your research, please consider citing the following paper.
@inproceedings{wang2021eip,
title={Elastic Tactile Simulation Towards Tactile-Visual Perception},
author={Wang, Yikai and Huang, Wenbing and Fang, Bin and Sun, Fuchun and Li, Chang},
booktitle={ACM International Conference on Multimedia (ACM MM)},
year={2021}
}
EIP is released under MIT License.