Denoise Restoration

Classic

Procedural API:
aydin.classic_denoise(noisy, *, batch_axes=None, chan_axes=None, variant=None)[source]

Method to denoise an image with Classic denoising restoration module.

Parameters
noisynumpy.ndarray

Image to denoise

batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

variantstr

Algorithm variant.

Returns
Denoised imagenumpy.ndarray
Object-Oriented API:
class aydin.restoration.denoise.classic.Classic(*, variant: Optional[str] = None, use_model=None, input_model_path=None, lower_level_args=None, it_transforms=None)[source]

Classic Image Denoising

static archive(source, destination)

Archives the model to given destination.

Parameters
sourcestr
destinationstr
static clean_model_folder(model_folder)

Method to clean model folder created

property configurable_arguments

Returns the configurable arguments that will be exposed on GUI and CLI.

denoise(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to denoise an image with trained Noise2Self FGR.

Parameters
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

noisy_imagenumpy.ndarray
Returns
responsenumpy.ndarray
get_translator()[source]

Returns the corresponding translator instance for given selections.

Parameters
feature_generatorFeatureGeneratorBase
regressorRegressorBase
Returns
itImageTranslatorBase
property implementations

Returns the list of discovered implementations for given method.

load(model_path: str)
Parameters
model_pathstr

whole path to the model including the model zip name

save(model_path)

Saves the latest trained model next to the input image file.

Parameters
model_pathstr
stop_running()[source]

Method to stop running N2S instance

train(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to run training for Noise2Self FGR.

Parameters
noisy_imagenumpy.ndarray
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

Returns
responsenumpy.ndarray

Noise2SelfFGR

Procedural API:
aydin.noise2self_fgr(noisy, *, batch_axes=None, chan_axes=None, variant=None)[source]

Method to denoise an image with trained Noise2Self FGR.

Parameters
noisynumpy.ndarray

Image to denoise

batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

variantstr

Algorithm variant.

Returns
Denoised imagenumpy.ndarray
Object-Oriented API:
class aydin.restoration.denoise.noise2selffgr.Noise2SelfFGR(*, variant: Optional[str] = None, use_model=None, input_model_path=None, lower_level_args=None, it_transforms=None)[source]

Noise2Self image denoising using the “Feature Generation & Regression” ( FGR) approach. Follows from the theory exposed in the <a href=”https://arxiv.org/abs/1901.11365”>Noise2Self paper</a>.

static archive(source, destination)

Archives the model to given destination.

Parameters
sourcestr
destinationstr
static clean_model_folder(model_folder)

Method to clean model folder created

property configurable_arguments

Returns the configurable arguments that will be exposed on GUI and CLI.

denoise(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to denoise an image with trained Noise2Self FGR.

Parameters
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

noisy_imagenumpy.ndarray
Returns
responsenumpy.ndarray
get_generator()[source]

Returns the corresponding generator instance for given selections.

Returns
generatorFeatureGeneratorBase
get_regressor()[source]

Returns the corresponding regressor instance for given selections.

Returns
regressorRegressorBase
get_translator(feature_generator, regressor)[source]

Returns the corresponding translator instance for given selections.

Parameters
feature_generatorFeatureGeneratorBase
regressorRegressorBase
Returns
itImageTranslatorBase
property implementations

Returns the list of discovered implementations for given method.

load(model_path: str)
Parameters
model_pathstr

whole path to the model including the model zip name

save(model_path)

Saves the latest trained model next to the input image file.

Parameters
model_pathstr
stop_running()[source]

Method to stop running N2S instance

train(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to run training for Noise2Self FGR.

Parameters
noisy_imagenumpy.ndarray
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

Returns
responsenumpy.ndarray

Noise2SelfCNN

Procedural API:
aydin.noise2self_cnn(image, *, batch_axes=None, chan_axes=None, variant=None)[source]

Method to denoise an image with Noise2Self CNN.

Parameters
imagenumpy.ndarray

Image to denoise

batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

variantstr

Algorithm variant.

Returns
Denoised imagenumpy.ndarray
Object-Oriented API:
class aydin.restoration.denoise.noise2selfcnn.Noise2SelfCNN(*, variant: Optional[str] = None, use_model=None, input_model_path=None, lower_level_args=None, it_transforms=None)[source]

Noise2Self image denoising using the “Convolutional Neural Networks” ( CNN) approach. Follows from the theory exposed in the <a href=”https://arxiv.org/abs/1901.11365”>Noise2Self paper</a>.

static archive(source, destination)

Archives the model to given destination.

Parameters
sourcestr
destinationstr
static clean_model_folder(model_folder)

Method to clean model folder created

property configurable_arguments

Returns the configurable arguments that will be exposed on GUI and CLI.

denoise(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to denoise an image with trained Noise2Self.

Parameters
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

noisy_imagenumpy.ndarray
Returns
responsenumpy.ndarray
get_translator()[source]

Returns the corresponding translator instance for given selections.

Returns
itImageTranslatorBase
property implementations

Returns the list of discovered implementations for given method.

load(model_path: str)
Parameters
model_pathstr

whole path to the model including the model zip name

save(model_path)

Saves the latest trained model next to the input image file.

Parameters
model_pathstr
stop_running()[source]

Method to stop running N2S instance

train(noisy_image, *, batch_axes=None, chan_axes=None, **kwargs)[source]

Method to run Noise2Self CNN training.

Parameters
noisy_imagenumpy.ArrayLike
batch_axesarray_like, optional

Indices of batch axes.

chan_axesarray_like, optional

Indices of channel axes.

Returns
responsenumpy.ArrayLike