Transforms¶

How do I transform my signal into another representation?

Features in the cobrabox.transforms domain. Access them as cb.<Feature> (canonical), or as cb.transforms.<Feature> / cb.feature.<Feature>.

Each feature has two forms: a class for building pipelines, and a one-shot function for a single call.

Extract analytic-signal representations along the time axis.

cb.analytic_signal(data, ...)  # one-shot
cb.AnalyticSignal(...).apply(data)  # composable, for pipelines

Tags: hilbert, instantaneous-phase, envelope, instantaneous-frequency, time-frequency, eeg, fmri, io:preserves-time

Real-valued FFT along the time axis.

cb.fourier_transform(data, ...)  # one-shot
cb.FourierTransform(...).apply(data)  # composable, for pipelines

Tags: fft, frequency-domain, eeg, fmri, io:frequency-output

Inverse of :class:~cobrabox.transforms.fourier_transform.FourierTransform.

cb.inverse_fourier_transform(data, ...)  # one-shot
cb.InverseFourierTransform(...).apply(data)  # composable, for pipelines

Tags: ifft, time-domain, io:time-output, req:frequency-input

Apply a notch filter to remove a specific frequency.

cb.notch_filter(data, ...)  # one-shot
cb.NotchFilter(...).apply(data)  # composable, for pipelines

Tags: filtering, preprocessing, notch, eeg, io:preserves-time