Real-Time Classification of Cardiac Events in Arrhythmia Disease

Millions of people lose their lives in a year due to the devastating heart condition known as Sudden Cardiac Arrest (SCA) due to Arrhythmia. Therefore, SCA could be predicted or diagnosed early, potentially preventing more deaths. In this study, we aim to devise a space and computation-time-efficient approach for the classification of SCA in near real-time. The MIT-BIH Arrhythmia dataset was used to train the classifiers such as CNN, hybrid of CNN and Long Short Term Memory (LSTM). The quantization technique was employed in the CNN models to achieve compression. To further improve compression, we proposed a method by integrating variable width quantization with retraining.