Algorithm development for real-time detection of premature ventricular contraction

Premature Ventricular Contraction (PVC) is one of the most common cardiac arrhythmias. PVC can occur in healthy people, but for those with frequent occurrence of PVCs, this can often be linked to pathological disorders of the heart. PVC detection allows the physician to diagnose heart disease accurately and also helps cardiac patients to be monitored effectively. This paper presents a novel algorithm for real-time PVC detection from ECG Lead II. Our methodology has low complexity in order to be applied to embedded devices. The developed algorithm is based on cardiac electrophysiology by considering 4 characteristics of ECG abnormalities, i.e. shorter RR-interval, wider QRS complex, changing of the QRS complex pattern and changing of the ST-level. The main parameters used in the algorithm are optimized to provide maximum performance of PVC detection. We tested the algorithm on 26 ECG records of MIT-BIH Arrhythmia Database. The performance of the proposed method has 97.75% of sensitivity and 98.80% of specificity. Furthermore, we also tested the algorithm on 16 selected records from Long-Term ST Database, with the results of 99.47% sensitivity and 99.24% specificity. The test results indicate that the algorithm presented in this work has high efficiency and high precision, which can be used to detect PVC for embedded devices in real-time.