glimpse.experiment.mf_wkmeans

GetMetaFeatures(prototypes)

Compute “meta-features for a set of prototypes.

Parameters:prototypes – Prototype data. Returns:Array of meta-features, with one row per prototype. Return type:array of float

TrainQualityModel(exp, num_regr_samples, pool, progress=None)

Train a patch quality model.

Parameters:

• num_regr_samples (int) – Number of patches to use for training the model.
• pool – Worker pool to use when extracting patches.
• progress – Handler for incremental progress updates.

Returns:

A trained model.

Return type:

sklearn.base.RegressorMixin

LearnPatchesFromImages(exp, num_regr_samples, num_samples, num_prototypes, pool, progress=None)

Learn patch models by meta-feature weighted k-Means clustering.

Weights are given by a feature quality prediction model using prototype “meta-features”.

Parameters:

• num_regr_samples (int) – Number of patches used to train quality prediction regression model.
• num_samples (int) – Number of samples used to cluster via k-Means.
• num_prototypes (int) – Number of centroids used for k-Means.
• pool – Worker pool to use when extracting patches.
• progress – Handler for incremental progress updates.

Returns:

Learned patches.

Return type:

array of float