deepmr.prox.LLRDenoiser#

class deepmr.prox.LLRDenoiser(*args: Any, **kwargs: Any)[source]#

Local Low Rank denoising.

The solution is available in closed-form, thus the denoiser is cheap to compute.

ndim#

Number of spatial dimensions.

Type:

int,

W#

Patch size (assume isotropic).

Type:

int, optional

ths#

Denoise threshold. The default is 0.1.

Type:

float, optional

trainable#

If True, threshold value is trainable, otherwise it is not. The default is False.

Type:

bool, optional

S#

Patch stride (assume isotropic). If not provided, use non-overlapping patches.

Type:

int, optional

rand_shift#

If True, randomly shift across spatial dimensions before denoising.

Type:

bool, optional

axis#

Axis assumed as coefficient axis (e.g., coils or contrasts). If not provided, use first axis to the left of spatial dimensions.

Type:

bool, optional

device#

Device on which the wavelet transform is computed. The default is None (infer from input).

Type:

str, optional

__init__(ndim, W, ths=0.1, trainable=False, S=None, rand_shift=True, axis=None, device=None)[source]#

Methods

__init__(ndim, W[, ths, trainable, S, ...])

forward(x)