Abstract: Objective: To develop an algorithm of signal quality assessment of ECG in the intensive care unit. Methods: The signal quality index (SQI) was obtained by analyzing the morphological and statistical characteristics of each waveform and their relationships. The heart rate was robustly estimated based on SQI. The algorithm was evaluated using more than 6?000 hours of simultaneously acquired ECG from a 437patient subset of the MultiParameter Intelligent Monitoring for Intensive Care II database and adding real ECG noises. Results: The SQI was decreased along with a decrease of the signal noise ratio. The SQI had high-positive correlation with the QRS detective sensitivity and positive predictivity. This method provided an accurate HR estimation even in the presence of high levels of persistent noises and artifacts. Conclusion: The signal quality assessment algorithm could provide an objective estimation of ECG quality.

Key words: Electrocardiogram, Noise, Signal quality assessment, Heart rate, Intensive care