florin.classification

Utilities for classifying connected components.

Classes

FlorinClassifier

Classification unit for weak object grouping.

Functions

classify

Classify a segmented object.

Functions

classify(obj, \*classes)

Multiclass classificaiton based on human-tuned boundaries.

Classes

FlorinClassifier(label, **kwargs)

Classify connected components based on boundary conditions.

class florin.classification.FlorinClassifier(label, **kwargs)

Classify connected components based on boundary conditions.

Parameters:

• label – The class label identifying this class. Can be any arbitrary label.
• boundaries – Pairs of values (2-tuples) passed as keyword arguments defining the boundaries to classify along. For example, passing area=(5, 10) tells this class that the objects it contains have an area/volume of 5 <= obj.area <= 10.

Methods

classify(self, obj)

Determine if an object is in this class.

classify(self, obj)

Determine if an object is in this class.

Parameters:obj (skimage.measure._regionprops.RegionProperties) – The object to classify. Returns:True if the object is within all defined boundaries else False. If no boundaries were provided, return True (e.g., the default class). Return type:bool

florin.classification.classify(obj, *classes)

Multiclass classificaiton based on human-tuned boundaries.

Parameters:

• obj (skimage.measure._regionprops.RegionProperties) – The object to classify.
• classes (florin.classify.FlorinClassifiers) – The classes to select from.

Returns:

Updates obj with a class label (obj.class_label) and passes it on for further processing.

Return type:

obj

Notes

In a typical FLoRIN pipeline, florin.reconstruct() will be called immediately after florin.classify().