This is a small libigl based code snippet that performs mesh to SDF collision
using several different schemes, including the optimization-based scheme from
Robust Optimization for Robust Signed Distance Function Collision.
The implementation is naive in that it's CPU based. If I decide to one day use this code beyond just as a learning exercise for myself, I'll probably implement this in CUDA. :)
To build, simply:
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ../
make -j8