Struct GraphWavelet 

pub struct GraphWavelet {
    pub scale: f64,
    /* private fields */
}

Graph wavelet based on the heat / diffusion kernel.

The diffusion kernel at scale t is K_t = U exp(−t Λ) Uᵀ. The wavelet at node s is the s-th column (or row) of K_t. This provides a spatially-localized, multi-scale representation of graph signals.

Reference

Hammond et al. (2011). “Wavelets on graphs via spectral graph theory.” Applied and Computational Harmonic Analysis, 30(2), 129–150.

Fields

scale: f64

Diffusion scale parameter t > 0.

Implementations

impl GraphWavelet

pub fn new(gft: &GraphFourierTransform, scale: f64) -> Result<Self>

Build the diffusion wavelet kernel at scale t from a GFT.

Arguments

• gft — precomputed graph Fourier transform.
• scale — diffusion scale t > 0.

pub fn apply(&self, signal: &Array1<f64>) -> Result<Array1<f64>>

Apply the wavelet kernel to a signal: y = K_t x.

pub fn wavelet_atom(&self, s: usize) -> Result<Array1<f64>>

Return the wavelet atom centered at node s (column s of K_t).

pub fn kernel(&self) -> &Array2<f64>

Return the full kernel matrix K_t (shape n × n).

Trait Implementations

impl Clone for GraphWavelet

fn clone(&self) -> GraphWavelet

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for GraphWavelet

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.