mhmohassan/ecg-analysis

ECG analysis to classify anterior myocardial infarction cases.

ECG Analysis

• data/scrape.py - Downloads the entire PTB dataset.
• data/patients.txt - A list of all the records.
• ecg_to_csv.py - Converts a final list of ECGs into CSV files.
• filter_align.py - Filter and determine one period of the ECG signal.
• separate.py - Split the data into positive cases and negative cases.
• tsne_plot.py - Extract coefficents from the waveforms and plot the t-SNE output.
• tsne.py - A Python implementation of t-SNE, written by the original author.

Requirements

wfdb, scipy.signal, numpy, pylab, pywt, matplotlib, glob

Running the Project

• Download Data - In the data folder, run scrape.py. In the home directory, run separate.py to get the positive and negative cases.
• Extract One Wave - Run filter_align.py once for positive and once for negative cases.
• Get Final CSV List - Run ecg_to_csv.py to get a final list of CSV files.
• t-SNE Plot - Run tsne_plot.py to extract coefficients from CSV files and get a final 2D t-SNE plot. Follow instructions in code to generate 3D plots.

Random Data

Unaligned ECG Signals

Aligned ECG Signals