spyrit.core.recon

Reconstruction methods and networks.

Classes

DCNet(noise, prep, sigma[, denoi])	Denoised completion reconstruction network.
Denoise_layer(std_dev_prior_or_size[, ...])	Defines a learnable Wiener filter that assumes additive white Gaussian noise.
LearnedPGD(noise, prep[, denoi, iter_stop, ...])	Learned Proximal Gradient Descent reconstruction network.
Pinv1Net(noise, prep[, denoi])	1D pseudo inverse reconstruction network.
PinvNet(noise, prep[, denoi])	Pseudo inverse reconstruction network.
PositiveParameters(params[, requires_grad])
PseudoInverse()	Moore-Penrose pseudoinverse.
TikhoNet(noise, prep, sigma[, denoi])	Tikhonov reconstruction network.
Tikhonov(meas_op, sigma[, approx])	Tikhonov regularization (aka as ridge regression).
TikhonovMeasurementPriorDiag(sigma, M)	Tikhonov regularisation with prior in the measurement domain.