dataeval.metrics.stats.hashstats#

dataeval.metrics.stats.hashstats(images, bboxes=None)#

Calculates hashes for each image

This function computes hashes from the images including exact hashes and perception-based hashes. These hash values can be used to determine if images are exact or near matches.

Parameters:

  • images (ArrayLike) – Images to hashing

  • bboxes (Iterable[ArrayLike] or None) – Bounding boxes in xyxy format for each image

Returns:

A dictionary-like object containing the computed hashes for each image.

Return type:

HashStatsOutput

See also

Duplicates

Examples

Calculating the statistics on the images, whose shape is (C, H, W)

>>> results = hashstats(stats_images)
>>> print(results.xxhash)
['6274f837b34ed9f0', '256504fdb6e3d2a4', '7dd0c56ca8474fb0', '50956ad4592f5bbc', '5ba2354079d42aa5']
>>> print(results.pchash)
['a666999999666666', 'e666999999266666', 'e666999966663299', 'e666999999266666', '96e91656e91616e9']