On applying multi-scale entropy to quality assessment of ECG collected via mobile phone

The paper used Multi-Scale Entropy (MSE) as a non-linear metric to determine whether mobile ECG could be used for clinical purpose, namely, quality assessment of ECG. This study firstly calculated the MSE value on the artificial ECG signals (i.e., the clean artificial ECG plus Gauss white noise, high frequency noise, low frequency noise and power-line noise, respectively), and analyzed the relationship between the MSE value and the content level of noise contained in the ECG signals. Besides, the optimum scale was also obtained. Secondly, this study verified the MSE metric for real ECGs developed from the PhysioNet/Computing in Cardiology Challenge 2011 (CINC 2011) of the MIT databases, and obtained classification accuracy including True positive (TP 44.211%) and True negative (TN 97.262%). Finally, the comparison among MSE, sample entropy (SampEn) and approximate entropy (ApEn) for quality assessment of ECG was given. The results indicated that the MSE was an effective metric for quality classification of ECG, and was superior to the other entropy approaches.