title: 3D computer vision and medical imaging algorithms in Python description: 73 algorithms in 19 families for point clouds, meshes and CT/MRI volumes, each with documented, data-tied defaults: registration (RANSAC, ICP, rigid, affine, B-spline, demons), reconstruction, segmentation, LiDAR, SDF, mesh and volume processing.

Algorithms¶

Nineteen families, 73 methods, 371 parameters, every default documented with the reason it was chosen and the data it was checked on. The methods are Open3D, PyVista/VTK, SimpleITK and SciPy code behind one call; pl.run("module.method", file, **params) validates the parameters before anything runs, and the high-level calls (align, segment, to_mesh, clean...) pick the method for the object type.

Start with Why our defaults work if you have been burnt by textbook parameters.

Family	Methods	Works on	High-level calls
Point Cloud Processing	`voxel_downsample`, `statistical_outlier_removal`, `radius_outlier_removal`, `mesh_to_pointcloud`	`.ply`, `.pcd`, `.txt`, `.xyz`	`points.downsample(voxel_size=...)`; `points.clean()`
Mesh Processing	`subdivision`, `simplification`, `mesh_to_pcd_uniform`, `mesh_to_pcd_poisson_disk`	`.ply`, `.obj`, `.stl`, `.off`	`mesh.decimate()`; `mesh.to_points(num_points=...)`
Voxelization & Volume	`voxelization`, `binary_erosion`	`.ply`, `.obj`, `.stl`, `.off`
Signed Distance Fields	`compute_sdf`, `occupancy_reconstruction`, `sdf_voxel_visualization`	`.ply`, `.obj`, `.stl`, `.off`
Surface Reconstruction	`poisson`, `convex_hull`, `alpha_shape`, `ball_pivoting`	`.ply`, `.pcd`, `.txt`, `.xyz`	`points.to_mesh()`
Point Cloud Registration	`ransac_global`, `icp_refinement`, `ransac_icp_pipeline`	`.ply`, `.pcd`, `.txt`, `.xyz`	`points.ransac(moving=...)`; `points.icp(moving=...)`; `target.align(moving)` on a `PointCloud`
Deformable Registration	`cpd_deformable`, `gaussian_deformation`	`.ply`, `.pcd`, `.txt`, `.xyz`
Point Cloud Segmentation	`ransac_plane`, `dbscan_clustering`	`.ply`, `.pcd`, `.txt`, `.xyz`	`points.segment(eps=..., min_points=...)`
Stereo & Depth Estimation	`depth_colorization`, `depth_filter`	`.ply`, `.pcd`, `.txt`, `.xyz`
LiDAR Scene Analysis	`height_colorization`, `ground_removal`, `object_clustering`	`.ply`, `.pcd`, `.txt`, `.xyz`, `.las`
LiDAR-Camera Fusion	`intensity_colorization`, `roi_extraction`, `multi_resolution`	`.ply`, `.pcd`, `.txt`, `.xyz`, `.las`
Features Extraction	`normal_estimation`, `fpfh_features`, `curvature_estimation`	`.ply`, `.pcd`, `.txt`, `.xyz`	`points.normals()`
Collision Detection & Measurement	`sdf_collision`, `penetration_depth`	`.ply`, `.obj`, `.stl`, `.off`
PyVista Mesh Processing	`pv_decimate`, `pv_subdivide`, `pv_smooth_laplacian`, `pv_smooth_taubin`, `pv_clean`, `pv_fill_holes`, `pv_compute_normals`, `pv_triangulate`, `pv_extract_surface`	`.ply`, `.obj`, `.stl`, `.vtk`, `.vtu`	`mesh.smooth()`; `mesh.clean()`
PyVista Point Cloud Processing	`pv_threshold`, `pv_clip_plane`, `pv_clip_box`, `pv_random_sample`, `pv_extract_points`	`.ply`, `.pcd`, `.vtk`, `.vtu`, `.obj`, `.stl`
PyVista Volume Processing	`pv_contour`, `pv_slice`, `pv_slice_orthogonal`, `pv_threshold_volume`, `pv_extract_geometry`, `pv_extract_cells`	`.vti`, `.vtk`, `.vtu`, `.ply`, `.obj`, `.stl`
PyVista Analysis	`pv_curvature`, `pv_distance`, `pv_glyph`, `pv_vector_magnitude`, `pv_scalar_stats`, `pv_extract_feature_edges`	`.ply`, `.obj`, `.stl`, `.vtk`, `.vtu`
SimpleITK Registration	`rigid`, `advanced_rigid`, `affine`, `bspline`, `demons`, `difference_analysis`	`.vti`, `.mha`, `.nrrd`, `.nii`, `.dcm`	`fixed.align(moving)` on a `Volume`
SimpleITK Segmentation	`otsu_threshold`, `manual_threshold`, `connected_components`, `watershed`	`.vti`, `.mha`, `.nrrd`, `.nii`	`volume.segment()`; `volume.segment(lower=..., upper=...)`

Models¶

Deep-learning models (PointNet, DINOv3 features, 3D JEPA, VAE anomaly detection, nnU-Net preparation) are documented under Models: reference scripts today, one-line pretrained access on the roadmap.