`imgtools.statistics`

Statistics functions

Links

Basic

`combine_mean_std`	Calculate the mean, standard deviation (std), and dataset size of the combination of two datasets.
`histogram`	Compute the histogram of an image.

Documents

imgtools.statistics.combine_mean_std(*stats: tuple[Tensor, Tensor, int]) → tuple[Tensor, Tensor, int]

Calculate the mean, standard deviation (std), and dataset size of the combination of two datasets. The

The function is present for evaluating the mean and std of a large dataset by computing its sub-datasets. To see the inference of the formula, check [1].

This function is not jit-able.

Parameters:

statstuple[torch.Tensor, torch.Tensor, int]: The [mean, standard deviation, number of samples] of dataset(s). np.ndarray type is also acceptable.

Returns:

torch.tensor: The mean value of the combined dataset.
torch.tensor: The standard deviation of the combined dataset.
int: The number of samples of the combined dataset.

References

[1] stack exchange - How do I combine standard deviations of two groups?: https://math.stackexchange.com/questions/2971315/how-do-i-combine-standard-deviations-of-two-groups

imgtools.statistics.histogram(img: Tensor, bins: int = 256, density: bool = False) → Tensor

Compute the histogram of an image.