Probability

class deeptrack.features.Probability(feature: Feature, probability: float | Callable[[...], float], *args: List[Any], **kwargs: Dict[str, any])

Bases: StructuralFeature

Resolve a feature with a certain probability

This feature conditionally applies a given feature to an input image based on a specified probability. A random number is sampled, and if it is less than the probability, the feature is resolved; otherwise, the input image remains unchanged.

Parameters

featureFeature

The feature to resolve conditionally.

probabilityfloat

The probability (between 0 and 1) of resolving the feature. A value of 0 ensures the feature is never resolved, while a value of 1 ensures it is always resolved.

*argsList[Any], optional

Positional arguments passed to the parent StructuralFeature class.

**kwargsDict[str, Any], optional

Additional keyword arguments passed to the parent StructuralFeature class.

Methods

get(image, feature, probability, random_number, **kwargs)

Resolves the feature if the sampled random number is less than the specified probability.

Example

In this example, the GaussianBlur is applied to the input image with a 50% chance.

>>> import numpy as np
>>> from deeptrack.features import Probability, GaussianBlur

Define a feature and wrap it with Probability:

>>> blur_feature = GaussianBlur(sigma=2)
>>> probabilistic_feature = Probability(blur_feature, probability=0.5)

Define an input image:

>>> input_image = np.ones((10, 10))

Apply the feature:

>>> output_image = probabilistic_feature(input_image)

Methods Summary

get(image, feature, probability, ...)

Resolve the feature if a random number is less than the probability.

Methods Documentation

get(image: ndarray, feature: Feature, probability: float, random_number: float, **kwargs)

Resolve the feature if a random number is less than the probability.

Parameters

imagenp.ndarray

The input image to process.

featureFeature

The feature to resolve conditionally.

probabilityfloat

The probability (between 0 and 1) of resolving the feature.

random_numberfloat

A random number sampled to determine whether to resolve the feature.

**kwargsDict[str, Any]

Additional arguments passed to the feature’s resolve method.

Returns

np.ndarray

The processed image. If the feature is resolved, this is the output of the feature; otherwise, it is the unchanged input image.