msdnet.validate module¶

Module for defining and processing validation sets.

class msdnet.validate.Validation[source]¶

Bases: abc.ABC

Base class for processing a validation set.

abstract validate(n)[source]¶

Compute validation metrics.

Parameters: n – network.Network to validate with
Returns: True if validation metric is lower than best validation error encountered, False otherwise.

abstract to_dict()[source]¶: Compute validation metrics.

abstract load_dict(dct)[source]¶

Return a dictionary containing all network variables and parameters.

Returns: all network variables and parameters

abstract classmethod from_dict(dct)[source]¶

Initialize Validation object from dictionary.

Parameters: dct – dictionary with all parameters

classmethod from_file(fn)[source]¶

Initialize Validation object from file.

Parameters: fn – filename

to_file(fn)[source]¶

Save all Validation object parameters to file.

Parameters: fn – filename

class msdnet.validate.MetricValidation(data, keep=True)[source]¶

Bases: msdnet.validate.Validation

Validation object that computes simple difference metrics.

Parameters

data – list of data.DataPoint objects to validate with.
keep – (optional) whether to keep the best, worst, and typical result in memory.

errorfunc(output, target, msk)[source]¶

Error function used for validation.

Parameters

output – network output image.
target – target image.
mask – mask image to indicate where to compute error function for.

Returns

error function value.

getbest()[source]¶

Return the input, target, and network output for best result.

Returns: list of images (input, target, network output)

getworst()[source]¶

Return the input, target, and network output for worst result.

Returns: list of images (input, target, network output)

getmedian()[source]¶

Return the input, target, and network output for median result.

Returns: list of images (input, target, network output)

validate(n)[source]¶

Compute validation metrics.

Parameters: n – network.Network to validate with
Returns: True if validation metric is lower than best validation error encountered, False otherwise.

to_dict()[source]¶: Compute validation metrics.

load_dict(dct)[source]¶

Return a dictionary containing all network variables and parameters.

Returns: all network variables and parameters

classmethod from_dict(dct)[source]¶

Initialize Validation object from dictionary.

Parameters: dct – dictionary with all parameters

class msdnet.validate.MSEValidation(data, keep=True)[source]¶

Bases: msdnet.validate.MetricValidation

Validation object that uses mean-squared error

errorfunc(output, target, msk)[source]¶

Error function used for validation.

Parameters

output – network output image.
target – target image.
mask – mask image to indicate where to compute error function for.

Returns

error function value.