Geometry processing library in pure Rust

Features

• Implicit/voxel/volume modeling
  • Voxel remeshing
  • Boolean operations
• Explicit modeling
  • Mesh simplification (decimation)
  • Isotropic remeshing

IO

Reading/writing mesh from/to STL file

You can read/write STL files using StlReader and StlWriter structs. Ony binary STLs are supported.

Example

use std::path::Path;

use baby_shark::{
    io::stl::{StlReader, StlWriter}, 
    mesh::corner_table::prelude::CornerTableF
};

fn main() {
    let mut reader = StlReader::new();
    let mesh: CornerTableF = reader.read_stl_from_file(Path::new("./read.stl"))
        .expect("Read mesh from STL file");

    let writer = StlWriter::new();
    writer.write_stl_to_file(&mesh, Path::new("./write.stl"))
        .expect("Save mesh to STL file");
}

Implicit modeling

Voxel remeshing

Voxel remeshing is a computational process used in computer graphics to reconstruct or optimize the topology of a three-dimensional (3D) model. Voxels are volumetric pixels that make up the 3D space, and remeshing involves reorganizing these voxels to create a more uniform and well-defined mesh structure. Also, it comes with the benefit of removing overlapping geometry, a valuable asset in sculpting applications.

Example

use baby_shark::{
    mesh::{builder, polygon_soup::data_structure::PolygonSoup},
    remeshing::voxel::VoxelRemesher,
};
use nalgebra::Vector3;

fn main() {
    let mesh: PolygonSoup<f32> = builder::cube(Vector3::zeros(), 1.0, 1.0, 1.0);
    let mut remesher = VoxelRemesher::default().with_voxel_size(0.02);
    let remeshed = remesher.remesh(&mesh).unwrap();
}

Boolean operations

Boolean operations are a set of operations that can be performed on volumes to combine or modify their shapes. The supported boolean operations in this library are:

• Union - combines two volumes into a single volume, resulting in a shape that includes the combined volume of both models.
• Subtract - removes the volume from another, resulting in a shape that is the difference between the two models.
• Intersect - returns the volume that is common to both models, resulting in a shape that includes only the overlapping region.

These boolean operations can be useful in various applications, such as creating complex shapes by combining simpler shapes, removing unwanted parts from a volume, or finding the intersection between two volumes. Supported boolean operations are:

Subtract	Union

Example

// See examples/boolean_operations.rs for full example
let mut bunny_volume = mesh_to_sdf.convert(&bunny_mesh).unwrap();

// Slice bunny with vertical boxes
let builder = VolumeBuilder::default().with_voxel_size(voxel_size);

for x in (-25..26).step_by(3) {
    let x = x as f32;
    let slice_box = builder.cuboid(Vec3::new(x, -20.0, 0.0), Vec3::new(x + 1.0, 20.0, 50.0));
    bunny_volume = bunny_volume.subtract(slice_box);
}

Explicit modeling

Isotropic remeshing

This algorithm incrementally performs simple operations such as edge splits, edge collapses, edge flips, and Laplacian smoothing. All the vertices of the remeshed patch are reprojected to the original surface to keep a good approximation of the input. Any of those operations can be turned off using appropriate method (with_<operation>(false)).

Example

let remesher = IncrementalRemesher::new()
    .with_iterations_count(10)
    .with_split_edges(true)
    .with_collapse_edges(true)
    .with_flip_edges(true)
    .with_shift_vertices(true)
    .with_project_vertices(true);
remesher.remesh(&mut mesh, 0.002f32);

Mesh simplification (decimation)

This library implements incremental edge decimation algorithm. On each iteration edge with lowest collapse cost is collapsed. Several stop condition are supported:

• Max error - algorithm stops when collapse lowest cost is bigger than given value
• Min faces count - algorithm stops when faces count drops below given value
• Bounding sphere - adaptive error algorithm based upon distance from a point. Useful for LOD mesh decimation.

Example

let mut decimator = EdgeDecimator::new()
    .decimation_criteria(ConstantErrorDecimationCriteria::new(0.0005))
    .min_faces_count(Some(10000));
decimator.decimate(&mut mesh);

Bounded Sphere Example

let origin = Point3::<f32>::origin();
let radii_error_map = vec![
    (10.0f32, 0.0001f32),
    (15.0f32, 0.05f32),
    (40.0f32, 0.8f32),
];

let criteria = BoundingSphereDecimationCriteria::new(origin, radii_error_map);

let mut decimator = EdgeDecimator::new().decimation_criteria(criteria);
decimator.decimate(&mut mesh);

2D triangulation

Triangulation2 struct implements fast 2D delaunay triangulation of points set.

Example

let mut triangulation = Triangulation2::new();
triangulation.triangulate(&vec![
    Point2::new(1.0, 2.0),
    Point2::new(5.0, 1.0),
    Point2::new(8.0, 6.0),
    Point2::new(2.0, 8.0)
]);

Constrained 2D triangulation

The ConstrainedTriangulation2 struct facilitates the constrained triangulation of a set of points. It builds upon the unconstrained Delaunay triangulation by inserting constrained edges. However, it's important to note that the resulting triangulation may not always be a Delaunay triangulation.

Conflicting constraints are automatically resolved through the following steps:

• When a new constrained edge intersects with another constrained edge, both edges are split into two at the intersection point
• When a new constrained edge intersects with a point, the edge is split into two at that point

Example

let points = vec![
    Point2::new(-3.0, 1.0),
    Point2::new(0.0, 0.0),
    Point2::new(0.0, 4.0),
    Point2::new(3.0, 2.0),
    Point2::new(6.0, 0.0),
    Point2::new(6.0, 4.0),
    Point2::new(9.0, 2.0)
];
let mut tri = ConstrainedTriangulation2::from_points(&points);
tri.insert_constrained_edge(0, 6);