nifreeze.data.splitting module

Data splitting helpers.

nifreeze.data.splitting.lovo_split(dataset, index, with_b0=False)

Produce one fold of LOVO (leave-one-volume-out).

Parameters:

• dataset (nifreeze.data.dmri.DWI) – DWI object

• index (int) – Index of the DWI orientation to be left out in this fold.

Returns:

• (train_data, train_gradients) (tuple) – Training DWI and corresponding gradients. Training data/gradients come from the updated dataset.

• (test_data, test_gradients) tuple – Test 3D map (one DWI orientation) and corresponding b-vector/value. The test data/gradient come from the original dataset.