Ventricular Tachycardia/Fibrillation Detection Algorithm for 24/7 Personal Wireless Heart Monitoring

摘要

This paper describes a Ventricular Tachycardia/Fibrillation (VT/VF) detection algorithm that is specifically designed for a 24/7 personal wireless heart monitoring system. This monitoring system uses Bluetooth enabled biosensors and smart phones to monitor continuously cardiac patients' vital signs. Our VT/VF algorithm is optimized for continuous real-time monitoring on smart phones with a high sensitivity and specificity. We studied and compared existing VT/VF algorithms and selected the one which suited best our requirements. However, we modified and improved the existing algorithm for the smart phone to achieve better performance results. We tested the algorithm on full-length signals from the physionet CU, MIT-db and MIT-vfdb databases [16] without any pre-selection of VT/VF or normal QRS-complex signals. We achieved 97% sensitivity, 98% accuracy and 98% specificity for our implementation which is excellent compared to existing algorithms.