Real-time Parallel Algorithmic System for Processing Multiple Data Simultaneously

Published

International Conference on Biomedical Engineering and Applications

Author

Sung-Jin Ahn, Jong-Doo Choi, Hoo-Hyun Kim, Kyung-Chul Kim, Hee-Seok Song

Abstract

In this study, we propose a system that analyzes multiple ECG data in real time within a single system. This system is designed with a parallel processing architecture, where algorithm modules perform different functions independently and operate autonomously. All algorithm modules can efficiently schedule data using index information, allowing for fast and accurate analysis of patient data without any bottlenecks. First, to verify the efficiency of the system, data was increased from 1 person to 300 people and compared. As a result, it was confirmed that while the existing system's processing speed linearly increased with the increase in data, the proposed system's processing speed increased non-linearly. Second, the proposed system was clinically studied on inpatients who underwent radiofrequency catheter ablation to verify the algorithm's performance. The results showed an arrhythmia detection performance with a sensitivity of 95.45% and a positive predictive value (PPV) of 83.09%. The proposed system has been shown to be effective and precise in processing large-scale patient data. It has the potential to enhance convenience and reduce costs for patients, while also enabling medical staff to respond promptly to emergencies through accurate arrhythmia alarms.