Struct MriCompressedSensing 

pub struct MriCompressedSensing {
    pub n: usize,
    pub m: usize,
}

Compressed sensing reconstruction for MRI-like k-space data.

In MRI, measurements are taken in Fourier (k-space) domain. This module provides a simplified 1-D model: the signal is sparse in the DCT domain, and k-space samples are random Fourier measurements.

Fields

n: usize

Signal length.

m: usize

Number of k-space samples (measurements).

Implementations

impl MriCompressedSensing

pub fn new(n: usize, m: usize) -> Self

Create an MRI-CS reconstruction problem for a signal of length n with m k-space measurements.

pub fn sample_kspace_indices(&self) -> Vec<usize>

Generate random k-space sampling indices in [0, n).

Returns m unique indices (or repeating if m > n).

pub fn build_measurement_matrix( &self, kspace_indices: &[usize], ) -> Vec<Vec<f64>>

Build a partial Fourier (cosine) measurement matrix from k-space indices.

Row i corresponds to k-space sample k_i; entry (i, j) is cos(2π k_i j / n) / sqrt(m).

pub fn reconstruct_fista( &self, measurements: &[f64], kspace_indices: &[usize], lambda: f64, max_iter: usize, ) -> Vec<f64>

Reconstruct a signal from k-space measurements using FISTA.

• measurements — observed k-space values
• kspace_indices — which k-space lines were sampled
• lambda — sparsity regularisation
• max_iter — number of FISTA iterations

Returns the reconstructed signal.

pub fn psnr(original: &[f64], reconstructed: &[f64], max_val: f64) -> f64

Compute the peak signal-to-noise ratio (PSNR) between two signals.

PSNR = 20 * log10(max_val / RMSE)