wiener

dnois.isp.wiener(signal: Tensor, kernel: Tensor, nsr: float | Tensor, ndim: int = 2) → Tensor

Wiener deconvolution:

\[\hat{x} = \mathcal{F}^{-1}\left\{\frac{h^*}{|h|^2 + \sigma^2}\mathcal{F}\{y\}\right\}\]

Parameters:

• signal (Tensor) – Signal to be deconvolved \(y\).

• kernel (Tensor) – Kernel used for deconvolution.

• nsr (float or Tensor) – Noise-to-signal ratio \(\sigma\).

• ndim (int) – Number of dimensions of the signal.

Returns:

Deconvolved signal.