Deep Recurrent Neural Networks for ECG Signal Denoising
share
We present a novel approach to denoise electrocardiographic signals (ECG), utilizing deep recurrent neural network built of Long-Short Term Memory (LSTM) units. The network is trained using synthetic data, generated using a dynamic model proposed by McSharry et al. as well as real data from Physionet PDB database of ECG signals. The results show that a 6-layer DRNN has a mean squared error as low as 0.0121 for denoising real signals with white noise of amplitude 0.2 mV, making it a viable alternative for other commonly used methods. We also investigate the impact of synthetic data on the network performance on real signals. Our findings show that networks trained with more synthetic data have better results than trained with more real data. We propose to explain this by means of the transfer learning framework and the analogy to human cognitive process.
READ FULL TEXT
