src.utils.util.prepare_stimulus_data

prepare_stimulus_data(all_data, data_subset, class_ignore_list, label_level)

Prepare data for modelling.

First filters data by class. Then rescales the image to [0-255] scale.

Parameters

• all_data – Dictionary of numpy arrays which contains all data.

• data_subset – A string value that denotes the train/test subset. Can only be “train” or “test”.

• class_ignore_list – A list of class labels to ignore.

• label_level – An integer that denotes the label hierarchy level. Lies between [0,3].

Returns

A tuple of 2d numpy arrays which contains input image and output class labels.

Return type

tuple

Raises

• ValueError – If label_level does not lie between [0, 3].

• ValueError – If data_subset does not contain either “train” or “test”.

• ValueError – If class_ignore_list is empty.

Return type

tuple