`pyheartlib.processing`

Module Contents

Classes

`Processing`	Methods for processing signals.
`STFT`	Short Time Fourier Transform.

class pyheartlib.processing.Processing

Methods for processing signals.

static apply(processors, signal)

Applies processors in order.

Parameters:

processors (list) – List of tuples of (‘processor_name’, params)
signal (numpy.ndarray) – Signal

static custom_processors(signal, processors=None)

Applies custom processors in order.

Parameters:

signal (numpy.ndarray) – one-dimensional signal.
processors (list) – Definition of processor functions as a list.

static remove_baseline(signal, sampling_rate=360)

Removes the signal baseline by applying two median filters.

Parameters:

signal (numpy.ndarray) – 1D ndarray.
sampling_rate (int, optional) – sampling frequency, by default 360

Returns:

Baseline removed signal.

Return type:

numpy.ndarray

static lowpass_filter_butter(signal, cutoff=45, sampling_rate=360, order=15)

Applies low pass filter to the signal.

Parameters:

signal (numpy.ndarray) – A one-dimensional signal.
cutoff (int, optional) – Filter parameter, by default 45
sampling_rate (int, optional) – Sampling frequency, by default 360
order (int, optional) – Filter parameter, by default 15

Returns:

Low pass filtered signal.

Return type:

numpy.ndarray

static denoise_signal(signal, remove_bl=True, lowpass=False, sampling_rate=360, cutoff=45, order=15)

Denoises the signal by removing the baseline wander and/or applying a low pass filter.

Parameters:

signal (numpy.ndarray) – A one-dimensional signal.
remove_bl (bool, optional) – If True, removes baseline wander, by default True
lowpass (bool, optional) – If True, applies a low pass filter, by default False
sampling_rate (int, optional) – Sampling frequency, by default 360
cutoff (int, optional) – Low pass filter parameter, by default 45
order (int, optional) – Low pass filter parameter, by default 15

Returns:

Denoised signal.

Return type:

numpy.ndarray

class pyheartlib.processing.STFT

Short Time Fourier Transform.

Example

>>> dpr = STFT()
>>> features = dpr.specgram(x, sampling_rate=360,
>>>                         nperseg=127, noverlap=122)

specgram(signals, sampling_rate=None, nperseg=None, noverlap=None)

Applies Short Time Fourier Transform on the signals.

Parameters:

signals (numpy.ndarray) – 2D array of raw signals. Each row is one signal.
sampling_rate (int, optional) – sampling_rate, by default None
nperseg (int, optional) – Window size (parameter of STFT), by default None
noverlap (int, optional) – Overlap (parameter of STFT), by default None

Returns:

3D array of transformed signals.

Return type:

numpy.ndarray

calc_feat_dim(samp, win, overlap)

Calculates the 2D spectral feature size.

Parameters:

samp (int) – Number of samples.
win (int) – Window size (parameter of STFT).
overlap (int) – Overlap (parameter of STFT).

Returns:

Height and width.

Return type:

int

pyheartlib.processing

Module Contents

Classes

`pyheartlib.processing`