Average

class deeptrack.optical.math.Average(axis: int | Callable[[...], int] = 0, features: list[Feature] | None = None, **kwargs: Any)

Bases: Feature

Average of input arrays.

Computes the mean of a list of arrays along the specified axis or axes. By default, averaging is performed along axis 0 (the batch dimension).

This operation is purely algebraic and does not interpret dimensions (e.g., spatial vs channel). All axes are treated uniformly and must be specified explicitly.

If features is provided, each feature is resolved first and the results are averaged.

Parameters

axis: int or tuple[int], optional

Axis or axes along which to compute the average. Defaults to 0.

features: list[Feature] or None, optional

List of features to resolve and average. Defaults to None.

Attributes

__distributed__: bool = False

Determines whether .get(…) is applied to each element independently (True) or to the list as a whole (False).

Methods

get(images, axis, **kwargs) -> np.ndarray | torch.Tensor

Computes the average of the input images along the given axis.

Examples

>>> import deeptrack as dt

Create two input images.

>>> import numpy as np
>>>
>>> input_image0 = np.ones((10, 30, 20)) * 2
>>> input_image1 = np.ones((10, 30, 20)) * 4

Define a pipeline with the average feature along the dimension 0.

>>> average = dt.Average(axis=0)
>>> output_image = average([input_image0, input_image1])
>>> output_image
(10, 30, 20)

Define a pipeline with the average feature along the dimension 1.

>>> average = dt.Average(axis=1)
>>> output_image = average([input_image0, input_image1])
>>> output_image.shape
(2, 30, 20)

Define a pipeline averaging each image.

>>> average = dt.Average(axis=(1, 2, 3))
>>> output_image = average([input_image0, input_image1])
>>> output_image.shape
(2,)

Methods Summary

get(images, axis, **kwargs)

Compute the average of input images along the specified axis(es).

Methods Documentation

get(images: list[ndarray | Tensor], axis: int | tuple[int], **kwargs: Any) → ndarray | Tensor

Compute the average of input images along the specified axis(es).

This method computes the average of the input images along the specified axis(es).

Parameters

images: array

The input images to average.

axis: int or tuple(int)

Axis or axes along which to average.

Returns

array

The average of the input images along the specified axis.