sdeconv.deconv.Spitfire

sdeconv.deconv.Spitfire#

class sdeconv.deconv.Spitfire(psf: Tensor, weight: float = 0.6, delta: float = 1, reg: float = 0.995, gradient_step: float = 0.01, precision: float = 1e-07, pad: int | tuple[int, int] | tuple[int, int, int] = 0)#

Variational deconvolution using the Spitfire algorithm

Parameters:

  • psf – Point spread function

  • weight – model weight between hessian and sparsity. Value is in ]0, 1[

  • delta – For 3D images resolution delta between xy and z

  • reg – Regularization weight. Value is in [0, 1]

  • gradient_step – Gradient descent step

  • precision – Stop criterion. Stop gradient descent when the loss decrease less than precision

  • pad – Image padding to avoid Fourier artefacts

__init__(psf: Tensor, weight: float = 0.6, delta: float = 1, reg: float = 0.995, gradient_step: float = 0.01, precision: float = 1e-07, pad: int | tuple[int, int] | tuple[int, int, int] = 0)#

Methods

__init__(psf[, weight, delta, reg, ...])

add_observer(observer)

Add an observer

adjoint_otf(psf)

Calculate the adjoint OTF of a PSF

notify(message)

Notify progress to observers

otf_3d(psf)

Calculate the OTF of a PSF

progress(value)

Notify progress to observers

run_2d(image)

Implements Spitfire for 2D images

run_3d(image)

Implements Spitfire for 3D images
Contents