ECG Monitoring and Anomaly Detection Based on Compressed Measurements

Abstract

Long-term monitoring systems based on wearable devices and local devices with computational capabilities -smartphone, smartwatch- could be used in the prevention of cardiovascular disease in risk subjects or during the follow-up for increasing the quality of life. In this paper we propose a lightweight solution that firstly exploits compressive sensing for locally reducing the amount of raw data, and successively employs a detection algorithm operating directly on the compressed domain for extracting only meaningful information to send at the medical staff. Performances of the proposed solution have been assessed under different conditions. Results show that the algorithm is able of identifying with a good precision and sensitivity the ECG features -QRS complexes and T, P waves- even with high compression ratios of about 20-50%.