cuvis_ai.transformation.reflectance.Reflectance

class cuvis_ai.transformation.reflectance.Reflectance(lower_bound: float = 0.0, upper_bound: float = 2.0)[source]

Bases: Node, BaseTransformation

Generic reflectance calculus: (data - dark) / (white - dark) Requires “Dark” and “White” references to be set in Metadata.

Parameters:

lower_bound (float, optional) – Threshold for the smallest allowed value, everything lower will be clamped. Set to None to allow any value. Default: 0.0
upper_bound (float, optional) – Threshold for the largest allowed value, everything higher will be clamped. Set to None to allow any value. Default: 2.0

__init__(lower_bound: float = 0.0, upper_bound: float = 2.0)[source]

Methods

`__init__`([lower_bound, upper_bound])
`check_input_dim`(X)	Check that the parameters for the input data data match user expectations
`check_output_dim`(X)	Check that the parameters for the output data data match user expectations
`forward`(X, references__White, references__Dark)	Apply reflectance calculus to the data.
`get_fit_requested_meta`()
`get_forward_requested_meta`()
`load`(params, serial_dir)	Load this node from a serialized graph.
`serialize`(serial_dir)	Serialize this node.
`set_fit_meta_request`(**kwargs)
`set_forward_meta_request`(**kwargs)

Attributes

`input_dim`	Returns the needed shape for the input data.
`output_dim`	Returns the shape for the output data.

check_input_dim(X)

Check that the parameters for the input data data match user expectations

Parameters: X (array-like): Input data.

Returns: (Bool) Valid data

check_output_dim(X)

Check that the parameters for the output data data match user expectations

Parameters: X (array-like): Input data.

Returns: (Bool) Valid data

forward(X: ndarray, references__White: ndarray, references__Dark: ndarray)[source]

Apply reflectance calculus to the data. Returns the data as percentage values between the “Dark” and “White” references set in the meta-data. e.g. A pixel value of 1.0 means that the pixel is as bright as the white reference at this pixel, 1.5 -> 50% brighter, 0.0 -> as bright as the dark reference, -0.2 -> 20% darker than the dark reference. The output values can be clamped by setting lower_bond and upper_bound.

Parameters

Xnp.ndarray: The input data array for which reflectance is computed. Must have the same shape as references__White and references__Dark.
references__Whitenp.ndarray: The white reference array. Defines the maximum intensity for each pixel.
references__Darknp.ndarray: The dark reference array. Defines the minimum intensity for each pixel.

np.ndarray: An array of reflectance values with the same shape as the input X. The reflectance values are computed as (X - references__Dark) / (references__White - references__Dark) and optionally clamped between the lower and upper bounds.

get_fit_requested_meta()

get_forward_requested_meta()

property input_dim: tuple[int, int, int]

Returns the needed shape for the input data. If a dimension is not important, it will return -1 in the specific position.

Returns: (tuple) needed shape for data

load(params: dict, serial_dir: str) → None[source]: Load this node from a serialized graph.

property output_dim: tuple[int, int, int]

Returns the shape for the output data. If a dimension is dependent on the input, it will return -1 in the specific position.

Returns: (tuple) expected output shape for data

serialize(serial_dir: str) → dict[source]: Serialize this node.

set_fit_meta_request(**kwargs)

set_forward_meta_request(**kwargs)