DENS-ECG: A deep learning approach for ECG signal delineation

With the technological advancements in field of tele-health monitoring, it is now possible to gather huge amount electro-physiological signals such as electrocardiogram (ECG). It therefore necessary develop models/algorithms that are capable analysing these massive data real-time. This paper proposes a deep learning model for real-time segmentation heartbeats. The proposed DENS-ECG algorithm, combines convolutional neural network (CNN) and long short-term memory (LSTM) detect onset, peak, offset different heartbeat waveforms P-waves, QRS complexes, T-waves, No waves (NW). Using ECG inputs, learns extract high level features through training process, which, unlike other classical machine based methods, eliminates feature engineering step. was trained validated on dataset with 105 records length 15 min each achieved an average sensitivity precision 97.95% 95.68%, respectively, using stratified 5-fold cross validation. Additionally, evaluated unseen examine its robustness detection, which resulted 99.61% 99.52%. empirical results show flexibility accuracy combined CNN-LSTM signal delineation. efficient easy use approach segmentation, could potentially be used monitoring systems.