TIBVH - Taichi-based Linear BVH

A high-performance Python package for 3D spatial data structures and geometry processing using Taichi for GPU acceleration.

Installation

Requirements

• Python 3.8+
• Taichi 1.6.0+
• NumPy 1.19.0+
• PyTorch 1.10.0+

Install from Source

git clone https://github.com/TATP-233/tibvh.git
cd tibvh
pip install -e .

Supported Geometry Types

• PLANE (0): Infinite planes with optional bounds
• SPHERE (2): Spheres defined by center and radius
• CAPSULE (3): Capsules (cylinders with hemispherical caps)
• ELLIPSOID (4): Ellipsoids with three different radii
• CYLINDER (5): Finite cylinders with caps
• BOX (6): Oriented bounding boxes
• TRIANGLES: Individual triangles for mesh processing

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Based on the Linear BVH construction algorithm by Tero Karras (NVIDIA Research)
• Powered by the Taichi programming language
• Inspired by modern GPU-accelerated spatial data structures

References

• Karras, T. (2012). "Maximizing Parallelism in the Construction of BVHs, Octrees, and k-d Trees". NVIDIA Research.
• Lauterbach, C. et al. (2009). "Fast BVH Construction on GPUs". Computer Graphics Forum.