grade.rate_distinctiveness

grade.rate_distinctiveness(features, match_threshold=0.6)

Grade images by their distinctiveness.

Compute the embedding recognizability score (ERS) for each image in the feature array.

Parameters

Name Type Description Default
features np.ndarray NumPy array of shape (image_count, feature_count) containing feature vectors for each image required
match_threshold float The threshold above which two images are considered a match. Must be between (0, 1) 0.6

Returns

Name Type Description
np.array Embedding recognizability score (ERS), a measure of distinctiveness, for every image in the dataset.