isoext 0.1.0

A collection of algorithms for iso-sufrace extraction on GPU. Supports pytorch.

isoext: Isosurface Extraction on GPU

Overview

Welcome to isoext — a Python library designed for efficient isosurface extraction, leveraging the power of GPU computing and comes with pytorch support. Our library attempts to implement a collection of classic isosurface extraction algorithms. Currently, only the following algorithms are supported, but more will come in the future:

• Marching cubes
  • lorensen: the original marching cubes algorithm from the paper Marching cubes: A high resolution 3D surface construction algorithm.

Installation

To install isoext, make sure CUDA Toolkit is installed and run:

pip install isoext

Quick Start

Here's a simple example to get you started:

import isoext
import torch 

def sphere_sdf(x):
    return x.norm(dim=-1) - 0.5

res = 128
x = torch.linspace(-1, 1, res)
y = torch.linspace(-1, 1, res)
z = torch.linspace(-1, 1, res)
grid = torch.stack(torch.meshgrid([x, y, z], indexing='xy'), dim=-1)
sdf = sphere_sdf(grid).cuda() # Only accept a gpu tensor from pytorch for now

aabb = [-1, -1, -1, 1, 1, 1]
isolevel = -0.2

v, f = isoext.marching_cubes(sdf, aabb, isolevel)
isoext.write_obj('sphere.obj', v, f)

Task List

• Fix docstring.
• Implement MC33.
• Add numpy support.

License

isoext is released under the MIT License. Feel free to use it in your projects.

Acknowledgments

• We use Thrust for GPU computing and nanobind for Python binding.
• The LUTs for lorensen are borrowed from Paul Bourke (https://paulbourke.net/geometry/polygonise/).