Simplify from Manifold about manifoldplus HOT 1 OPEN

Also using ManifoldPlus on shapenet as part of a SDF generation pipeline. Generally I see few artifacts at --depth 9, however there are still problems with complex objects like cars, usually when there are thinly connected parts, ManifoldPlus seems to fill in those regions.

For my purposes, I only need correctly oriented faces and distance is computed using the original mesh, so my tolerance for artifacts is higher.

Here is a snippet to batch process shapenet objects:

from tqdm import tqdm

def fixmesh(paths):
    input_path, output_path = paths
    try:
        completed = subprocess.run(["/home/aluo/Tools2/ManifoldPlus/build/manifold", "--input", input_path, "--output", output_path, "--depth", "9"], timeout=60, check=True, capture_output=True)
    except:
        print("{} failed to run",format(input_path))
        return False
    return True

#our_paths = list(zip(inputs, outputs))
Parallel(n_jobs=4)(delayed(fixmesh)(i) for i in tqdm(our_paths))

Edit: For shapenetv2 specifically, I have also tried outer hull algorithms, but these generally fail by causing thin structures to disappear. Mesh -> Voxel -> Mesh methods generally fail due to the non watertight nature of a lot of the meshes.

from manifoldplus.