braindecode.augmentation.CutMix#

class braindecode.augmentation.CutMix(alpha, beta_per_sample=False, random_state=None)[source]#

Implements CutMix for EEG data.

A random time region of each example is replaced by the same region of a shuffled example of the batch, see [1]. As in the reference code [2], a coefficient lam is drawn from a Beta distribution, the region is (1 - lam) * n_times long (the 1-D analogue of the cut area ratio 1 - lam) around a uniformly drawn center, it is clipped at the trial boundaries, and lam is then recomputed from the pasted length.

Like Mixup, it applies to whole batches and returns targets (y_a, y_b, lam) for braindecode.training.mixup_criterion().

Parameters:

  • alpha (float) – Parameter of the Beta distribution of lam. alpha <= 0 cuts nothing.

  • beta_per_sample (bool (default=False)) – By default, one lam per batch is drawn. If True, one per sample.

  • random_state (int | numpy.random.Generator, optional) – Seed to be used to instantiate numpy random number generator instance. Defaults to None.

References

[1]

Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, Youngjoon Yoo (2019). CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV) Online: https://arxiv.org/abs/1905.04899

Methods

get_augmentation_params(*batch)[source]#

Return idx_perm and the cut_t_start, cut_t_len of each region.

static operation(X, y, idx_perm, cut_t_start, cut_t_len)[source]#

Paste a time region of shuffled examples into the original examples.

See [1] for details.

Parameters:

  • X (Tensor) – EEG data in form batch_size, n_channels, n_times

  • y (Tensor) – Target of length batch_size

  • idx_perm (Tensor) – Permuted indices of example that are mixed into original examples.

  • cut_t_start (Tensor) – Start index of the pasted time region for each example.

  • cut_t_len (Tensor) – Length of the pasted time region for each example.

Returns:

X, y. Where X is augmented and y is a tuple of length 3 containing the labels of the two mixed examples and the fraction of each example kept from the original, 1 - cut_t_len / n_times.

Return type:

tuple[Tensor, Tensor]

References

[1]

Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, Youngjoon Yoo (2019). CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features. In 2019 IEEE/CVF International Conference on Computer Vision (ICCV) Online: https://arxiv.org/abs/1905.04899