darsia.LabelColorPathMapRegression#

class LabelColorPathMapRegression(labels, color_range, resolution=11, mask=None, ignore_labels=[])[source]#

Bases: object

Algorithmic definition of relative color paths through regression for labeled images.

Methods

`base_color_image`(image)	Create an image where each label is colored by its base color.
`expand_color_spectrum`(color_spectrum[, ...])	Expand the color spectrum through linear regression.
`find_color_path`(color_spectrum[, ignore, ...])	Find relative color paths for each label in the spectrum map.
`get_base_colors`(image[, verbose])	Get the base colors for each label in the image.
`get_color_spectrum`(images[, baseline, ...])	Get the color spectrum for each label in the image.
`get_mean_base_color`(image)	Get the mean base color across all labels in the image.

Attributes

`labels`	Labeled image to define color paths for.
`color_range`	Color range for the regression.
`discrete_color_range`	Discrete color range for regression.
`mask`	Mask to apply when analyzing colors.
`ignore_labels`	Labels to ignore during color path regression.
`color_mode`	Color mode.

base_color_image(image)[source]#

Create an image where each label is colored by its base color.

Parameters:

imagedarsia.Image: The image from which to extract base colors.

Returns:

darsia.Image: An image where each label is colored by its base color.

expand_color_spectrum(color_spectrum, min_points=6, verbose=False)[source]#

Expand the color spectrum through linear regression.

Parameters:

color_spectrumColorSpectrum or LabelColorSpectrumMap: The color spectrum to expand.
min_pointsint: Minimum number of significant points to perform regression.
verbosebool: Whether to print additional information.

Returns:

LabelColorSpectrumMap: The expanded color spectrum.

find_color_path(color_spectrum, ignore=None, num_segments=1, directory=None, weighting='threshold', mode='auto', preview_image=None, preview_images=None, preview_baseline=None, ignore_labels=None, verbose=False)[source]#

Find relative color paths for each label in the spectrum map.

Parameters:

color_spectrumLabelColorSpectrumMap: The color spectrum map to analyze.
ignoreLabelColorSpectrumMap | None: The color spectrum map to ignore.
num_segmentsint: The number of segments for the color path.
weightingLiteral: How to use histogram counts when fitting each path. "threshold" (default) uses binary 0/1 weights from the spectrum threshold – counts are not used beyond that. "wls", "wls_sqrt", and "wls_log" use count-weighted average errors – see _find_color_path() for details.
modeLiteral: Color-path selection mode. "auto" returns the automated result. "manual" enables interactive key-color postprocessing.
preview_imagedarsia.Image | None: Calibration image for manual preview.
preview_imageslist[darsia.Image] | None: Calibration images for manual preview navigation.
preview_baselinedarsia.Image | None: Baseline used for relative color preview.
ignore_labelslist[int] | None: Label IDs to skip fitting and produce zero-paths for instead. Defaults to None (no labels ignored).
verbosebool: Whether to print additional information.

Returns:

LabelColorPathMap: The relative color path map through the significant boxes.

get_base_colors(image, verbose=False)[source]#

Get the base colors for each label in the image.

Parameters:

imagedarsia.Image: The image from which to extract base colors.
verbosebool: Whether to print additional information.

Returns:

LabelColorMap: An object containing the base colors for each label.

get_color_spectrum(images, baseline=None, ignore=None, threshold_zero=0.0, threshold_significant=0.0, path=None, verbose=False)[source]#

Get the color spectrum for each label in the image.

The color spectrum is calculated by analyzing the distribution of colors within each label across the provided images. This involves creating a 3D histogram of color occurrences, which is then normalized to identify significant colors.

Parameters:

imageslist[darsia.Image]: The images to analyze.
baselinedarsia.Image | None: The baseline image for comparison.
ignoredarsia.LabelColorSpectrumMap | darsia.ColorSpectrum | None: Colors to ignore in the spectrum.
threshold_zerofloat: The threshold for zeroing out insignificant colors.
threshold_significantfloat: The threshold for significant colors.
pathPath: Folder to store images.
verbosebool: Whether to print verbose output.

Returns:

LabelColorSpectrumMap: The color spectrum for each label.

get_mean_base_color(image)[source]#

Get the mean base color across all labels in the image.

Parameters:

imagedarsia.Image: The image from which to extract base colors.

Returns:

np.ndarray: The mean base color across all labels.

color_mode#: Color mode.

color_range#: Color range for the regression.

discrete_color_range#: Discrete color range for regression.

ignore_labels#: Labels to ignore during color path regression.

labels#: Labeled image to define color paths for.

mask#: Mask to apply when analyzing colors.