This is the official implementation of A Time-Dependent Inclusion-Based Method for Continuous Collision Detection between Parametric Surfaces (accepted as SIGGRAPH Asia 2024 Journal Paper).
We have provided the core code for CCD between parametric surfaces in the ./core folder, including implementation of geometric primitives, CCD solvers, and some auxiliary functions. Different test scenarios are set in the ./scene folder.
Also, we have provided a denegerate version of CCD methods for EE tests and VF tests between linear triangle meshes in the ./benchmark folder of the benchmark branch.
We have implemented the traditional inclusion-based method (described in Section 3) and our proposed time-dependent inclusion-based method (described in Section 4). Both methods can adopt AABB or OBB as the inclusion type.
CCD methods between parametric surfaces have been tested on different geometric primitives, including triangular patches and rectangular patches of 1st/2nd/3rd order. Each patch is represented by a list of control points. Taking third-order patches as an example, we illustrate the order below.
- Each control point of an rectangular patch of order
$n$ is denoted by 2 integers indicating its position on the$(u,v)$ grid. - Each control point of an triangular patch of order
$n$ is denoted by 3 integers, which add up to$n$ , indicating its position on the$(u,v,w)$ grid.
This repository has contained 3 experiments for CCD between parametric surfaces.
- Single test. We manually design a simple case to show how to use the CCD solver.
- Random test. It randomly generates a number of CCD cases between third-order rectangular patches and report the average time consumption. The geometric primitive can be changed by editing the template parameters in the
mainfunction. - Bunny-torus test (Figure 2). CCD is applied between 2 multi-patch objects. Both the bunny and the torus undergo imitated rigid-body motions.
Additionally, it includes one experiment for EE/VF tests in the benchmark branch. This experiment evaluates CCD solvers on the large-scale benchmark released by CCD-Wrapper.
- Install xmake.
- Run the following lines to install this project:
git clone https://github.com/xw-c/TDIB-CCD.git
cd TDIB-CCD
xmake
xmake run scene
Optional arguments:
-s Type of CCD solver. Use "trad" for the traditional method, or "td" for our time-dependent method.
-e Type of experiment. Use "rand" for the random test, "single" for the single test, or "bunny" for the bunny-torus test.
-b Type of inclusions. Use "aabb" for AABB, or "obb" for OBB.
-d Precision requirement, i.e., maximum width of the parametric interval.
-k Number of generated cases. Only used in the "rand" experiment.
The default setting is
xmake run scene -s td -e rand -b obb -d 1e-5 -k 100
Put the ./scene/bunny292.obj file under the same folder as scene.exe before you run the bunny-torus test.
To build the experiment for EE/VF tests, checkout to the benchmark branch, follow the instructions on CCD-Wrapper, and use similar commands to install and run it.
Feel free to open a github issue or contact us at [email protected] / [email protected].
Please consider citing our paper if you find it useful for your project :)
@article{10.1145/3687960,
author = {Chen, Xuwen and Yu, Cheng and Ni, Xingyu and Chu, Mengyu and Wang, Bin and Chen, Baoquan},
title = {A Time-Dependent Inclusion-Based Method for Continuous Collision Detection between Parametric Surfaces},
year = {2024},
issue_date = {December 2024},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
volume = {43},
number = {6},
issn = {0730-0301},
url = {https://doi.org/10.1145/3687960},
doi = {10.1145/3687960},
journal = {ACM Trans. Graph.},
month = nov,
articleno = {223},
numpages = {11}
}