A hybrid deep learning approach incorporating convolutional and recurrent neural networks classified cardiac rhythms in the MIT-BIH dataset with an average accuracy of 98%.
Does a hybrid deep learning approach accurately classify cardiac beats in ECG signals?
A hybrid deep learning approach combining convolutional and recurrent neural networks achieved 98% average accuracy in classifying cardiac rhythms on the MIT-BIH dataset.
Electrocardiogram (ECG) signal is used recent days to identify the abnormalities related to heart beat of human and to identify the functionality of cardiovascular system. Now a days this ECG signal is very much useful to classify the heartbeats. Several researchers have shown much interest in this area, but still much more analysis is still need to classify data more accurately. This paper has introduced a hybrid approach for deep learning by incorporating both the revolution and a recurrent deep neural network to reliably identify the cardiac beats by using five rhythms. The method on the MIT-BIH Diagnostics dataset of PhysionNets has been tested. According to results, the proposed method will classify the data set in the rhythmic classification by an average of 98 percent.
- et al. (Thu,) conducted a other in Cardiac rhythm abnormalities. Hybrid deep learning approach (convolutional and recurrent neural network) was evaluated on Rhythmic classification accuracy. A hybrid deep learning approach incorporating convolutional and recurrent neural networks classified cardiac rhythms in the MIT-BIH dataset with an average accuracy of 98%.