(With inputs from Karan Chadha and Abhin Shah).

Last semester, under the Digital Signal Processing course, we worked on an interesting application project. Our goal was to analyse ECG signals using DSP techniques and identify heart attacks (more precisely, Anterior Myocardial Infractions). We have hosted the project on Github.

Literature Survey

Data Collection

Data was collected from a standard ECG analysis database called Physikalisch-Technische Bundesanstalt (PTB). We scraped Physionet using a python script. Each ECG waveform is accompanied with a header file with details on the diagnosis. We extracted all Anterior Myocardial Infarction cases in positive.txt and healthy cases in control.txt.

Pre-Processing

Extracting Coefficients

Results