iris.nodes.binarization package

Submodules

iris.nodes.binarization.multilabel_binarization module

class iris.nodes.binarization.multilabel_binarization.MultilabelSegmentationBinarization(eyeball_threshold: float = 0.5, iris_threshold: float = 0.5, pupil_threshold: float = 0.5, eyelashes_threshold: float = 0.5, callbacks: List[Callback] = [])

Bases: Algorithm

Implementation of a binarization algorithm for multilabel segmentation. Algorithm performs thresholding of each prediction’s channel separately to create rasters based on specified by the user classes’ thresholds.

class Parameters(*, eyeball_threshold: ConstrainedFloatValue, iris_threshold: ConstrainedFloatValue, pupil_threshold: ConstrainedFloatValue, eyelashes_threshold: ConstrainedFloatValue)

Bases: Parameters

Parameters class for MultilabelSegmentationBinarization objects.

eyeball_threshold: float

eyelashes_threshold: float

iris_threshold: float

pupil_threshold: float

run(segmentation_map: SegmentationMap) → Tuple[GeometryMask, NoiseMask]

Perform segmentation binarization.

Parameters:

segmentation_map (SegmentationMap) – Predictions.

Returns:

Binarized geometry mask and noise mask.

Return type:

Tuple[GeometryMask, NoiseMask]

iris.nodes.binarization.specular_reflection_detection module

class iris.nodes.binarization.specular_reflection_detection.SpecularReflectionDetection(reflection_threshold: int = 254)

Bases: Algorithm

Apply a threshold to the IR Image to detect specular reflections.

class Parameters(*, reflection_threshold: ConstrainedIntValue)

Bases: Parameters

Parameter class for FusedSemanticSegmentation class.

reflection_threshold: int

run(ir_image: IRImage) → NoiseMask

Thresholds an IRImage to detect Specular Reflection.

Parameters:

ir_image (IRImage) – Infrared image object.

Returns:

a binary map of the thresholded IRImage.

Return type:

NoiseMask

Module contents