colour.recovery.Dataset_Otsu2018#

class colour.recovery.Dataset_Otsu2018(shape: SpectralShape | None = None, basis_functions: NDArrayFloat | None = None, means: NDArrayFloat | None = None, selector_array: NDArrayFloat | None = None)[source]#

Store all the information needed for the Otsu et al. (2018) spectral upsampling method.

Datasets can be either generated and converted as a colour.recovery.Dataset_Otsu2018 class instance using the colour.recovery.Tree_Otsu2018.to_dataset() method or alternatively, loaded from disk with the colour.recovery.Dataset_Otsu2018.read() method.

Parameters:

  • shape (SpectralShape | None) – Shape of the spectral data.

  • basis_functions (NDArrayFloat | None) – Three basis functions for every cluster.

  • means (NDArrayFloat | None) – Mean for every cluster.

  • selector_array (NDArrayFloat | None) – Array describing how to select the appropriate cluster. See colour.recovery.Dataset_Otsu2018.select() method for details.

Attributes

  • shape

  • basis_functions

  • means

  • selector_array

Methods

References

[OYH18]

Examples

>>> import os
>>> import colour
>>> from colour.characterisation import SDS_COLOURCHECKERS
>>> from colour.colorimetry import sds_and_msds_to_msds
>>> reflectances = sds_and_msds_to_msds(
...     SDS_COLOURCHECKERS["ColorChecker N Ohta"].values()
... )
>>> node_tree = Tree_Otsu2018(reflectances)
>>> node_tree.optimise(iterations=2, print_callable=lambda x: x)
>>> dataset = node_tree.to_dataset()
>>> path = os.path.join(
...     colour.__path__[0],
...     "recovery",
...     "tests",
...     "resources",
...     "ColorChecker_Otsu2018.npz",
... )
>>> dataset.write(path)  
>>> dataset = Dataset_Otsu2018()  
>>> dataset.read(path)
__init__(shape: SpectralShape | None = None, basis_functions: NDArrayFloat | None = None, means: NDArrayFloat | None = None, selector_array: NDArrayFloat | None = None) → None[source]#
Parameters:

  • shape (SpectralShape | None) –

  • basis_functions (NDArrayFloat | None) –

  • means (NDArrayFloat | None) –

  • selector_array (NDArrayFloat | None) –

Return type:

None

Methods

__init__([shape, basis_functions, means, ...])

cluster(xy)

Return the basis functions and dataset mean for the given CIE xy coordinates.

read(path)

Read and loads a dataset from an .npz file.

select(xy)

Return the cluster index appropriate for the given CIE xy coordinates.

write(path)

Write the dataset to an .npz file at given path.

Attributes

basis_functions

Getter property for the basis functions of the Otsu et al. (2018) dataset.

means

Getter property for means of the Otsu et al. (2018) dataset.

selector_array

Getter property for the selector array of the Otsu et al. (2018) dataset.

shape

Getter property for the shape used by the Otsu et al. (2018) dataset.