sxqslim : Self-intersection free mesh simplification with QEM

sxqslim is a Python package for mesh simplification that ensures the output mesh is manifold and free of self-intersections. It leverages Quadric Error Metrics (QEM) for simplification and employs an AABB tree for efficient self-intersection detection during each iteration.

• Utilizes libigl and an AABB tree for spatial indexing (https://github.com/lohedges/aabbcc)
• Adds self-intersection tests using the AABB tree for every iteration
• input: Manifold, self-intersection free mesh
• Output: Decimated manifold, self-intersection free mesh

Compile

cd qslim
git submodule add https://github.com/libigl/libigl.git external/libigl
git submodule update --init --recursive
pip install -e .

Simplifying a Mesh

1. Load your mesh using a compatible mesh loader, such as trimesh.
2. Initialize the MeshSimplify object with the vertices and faces of your mesh. This step initializes various data structures required for the simplification process, including the AABB tree for spatial indexing and edge data for connectivity analysis.
3. Call the process method on the MeshSimplify object, specifying the simplification ratio.
4. Use the simplified vertices and faces to create a new mesh.

Example code

import sxqslim
import trimesh

file_path = './mesh/bunny.obj'

mesh = trimesh.load(file_path)
vertices = mesh.vertices
faces = mesh.faces

# Initialize the MeshSimplify object
mesh_simplify = sxqslim.MeshSimplify(vertices, faces)

# Perform the mesh simplification process
verts, faces = mesh_simplify.process(ratio=0.01)

simplified_mesh = trimesh.Trimesh(vertices=verts, faces=faces)

output_file_path = './simplified_mesh.obj'
simplified_mesh.export(output_file_path)

print(f"Mesh saved to {output_file_path}")

Run the example code with:

python example.py

Setting the ratio to 0.3 retains only 30% of the original vertices.

Reference

Surface Simplification Using Quadric Error Metrics, 1997
https://www.cs.cmu.edu/~./garland/Papers/quadrics.pdf