This work was presented at the Physionet Challenge 2017 presented at the Computing in Cardiology conference 2017. The Challenge consisted of classifying short single-lead ECG segments with 10-60 seconds duration into one of the following classes:
|N||normal sinus rhythm|
|A||atrial fibrillation (AF)|
|O||other cardiac rhythms|
Two methodologies are proposed and described in distict forlder within this repo:
- Classic feature-based MATLAB approach (
- Deep Convolutional Network Approach in Python (
The complete code is available on my Github .
When using this code, please cite our paper :
Andreotti, F., Carr, O., Pimentel, M.A.F., Mahdi, A., & De Vos, M. (2017). Comparing Feature Based Classifiers and Convolutional Neural Networks to Detect Arrhythmia from Short Segments of ECG. In Computing in Cardiology. Rennes (France).