Automatic detection for multi-labelled cardiac arrhythmia based on frame blocking pre-processing and Residual Networks

Requirements The code is implemented based on python 3.6 with tensorflow==2.0.0 and keras 2.3.1. Please check Requirements.txt.

Scripts: Preprocess.py: scripts for loading data and preprocessing. Proposed_model.py: scripts for traing, validation and testing on the proposed model. CNN_LSTM.py: scripts for traing, validation and testing on the plain CNN+LSTM model. CNN_attenBILSTM.py: scripts for traing, validation and testing on the plain CNN with attention-based BiLSTM model. Challengebest_model.py: scripts for traing, validation and testing on the Challenge-best model on CPSC 2018. Bar.py: scripts for plotting of F1 scores. ROC.py: scripts for plotting of ROC curves.