Struct RbfDeformer 

pub struct RbfDeformer { /* private fields */ }

A Radial Basis Function (RBF) deformer for 3D point transformations.

This struct implements a deformable model that can smoothly interpolate between a set of control points in 3D space. It’s particularly useful for mesh deformation, shape morphing, and other spatial transformations.

Fields

• x_mean: Mean of the input control points for normalization
• x_std: Standard deviation of the input control points for normalization
• y_mean: Mean of the output control points for denormalization
• y_std: Standard deviation of the output control points for denormalization
• removed_columns: Indices of dimensions with zero variance in the output
• rbf: The underlying RBF interpolator

Implementations

impl RbfDeformer

pub fn new( x: Vec<[f64; 3]>, y: Vec<[f64; 3]>, kernel_name: Option<&str>, epsilon: Option<f64>, ) -> Result<Self>

Creates a new RbfDeformer instance.

Arguments

• x - Input control points (n×3 array)
• y - Corresponding output control points (n×3 array)
• kernel_name - Name of the kernel function to use (optional, defaults to “gaussian”):
  • “linear”: Linear kernel
  • “cubic”: Cubic kernel
  • “gaussian”: Gaussian kernel (default)
  • “multiquadric”: Multiquadric kernel
  • “inverse_multiquadratic”: Inverse multiquadric kernel
  • “thin_plate_spline”: Thin plate spline kernel
• epsilon - Bandwidth parameter for the kernel (optional, defaults to 1.0)

Returns

A new RbfDeformer instance or an error if creation fails.

pub fn deform(&self, points: &[[f64; 3]]) -> Result<Vec<[f64; 3]>>

Deforms input points using the learned RBF transformation.

Arguments

• points - A slice of 3D points to transform

Returns

A Vec of transformed points with the same length as the input, or an error string if the transformation fails.

Example

let points = [[0.5, 0.5, 0.5], [0.2, 0.8, 0.4]];
let deformed = deformer.deform(&points).unwrap();
assert_eq!(deformed.len(), 2);