sidesh27 / pysigpro

Python package for processing currently ECG and EEG signals

PySigPro

Description

PySigPro is a work in progress one-stop comprehensive Python package that serves as a feature extraction tool which extracts features from various domains. These features are frequently used in algorithms for ECG/EEG signal processing. Powerful feature extraction system is necessary as features provide useful and relative information that aids algorithms in their performance. The raw signals can be split in segments of time, termed as a window, from which features can be extracted.

ECG FEATURES

Time domain features

• MEAN_NNI, SDNN, SDSD, NNI_20, PNNI_20, NNI_50, PNNI_50, RMSSD, MEDIAN_NNI, CVSD, CVNNI, MEAN_HR, MAX_HR, MIN_HR, STD_HR

Geometrical features

• TRIANGULAR_INDEX

Entropy features

• SAMPLE_ENTROPY

EEG FEATURES

Time domain features

• Mean, Standard deviation, Kurtosis, Skewness, Variance, Mode, Median, Minimum, Maximum, Coefficient of variation, Energy, Average power, RMS value, Line length, Non linear energy, Hjorth parameters : Activity, Mobility, Complexity

Entropy domain features

• Shannon entropy, Approximate entropy, Sample entropy, Permutation entropy, Singular value decomposition, Higuchi fractal dimension, Number of zero crossings

Requirements

• pandas
• numpy
• numba
• scipy
• scikit-learn
• nolds

    $ pip install -r requirements.txt

Installation

    $ pip install git+https://github.com/vishaln15/pysigpro.git

You can also clone the repository:

    $ git clone https://github.com/vishaln15/pysigpro.git
    $ pip install -r requirements.txt
    $ python setup.py install

Authors

Vishal Nagarajan

Ashwini Muralidharan

Sideshwar J B