A knowledge-based real time embedded platform for arrhythmia beat classification

Abstract

This paper aims for accurate classification of ECG beats in real-time. The design, implementation and results of FPGA oriented SOC based embedded platform in laboratory experimental setup is presented. The platform is designed using high-level systematic approach having the capability of real-time classification of ECG beats at the place of patients with high accuracy which will be helpful to enhance the healthcare for cardiovascular diseases. The algorithm involves the integration of the R-peak detection algorithm and Rulebased approach to classify four generic heartbeat classes namely Normal, PVC, Ventricular Fibrillation/Ventricular Flutter, 2° heart block beat. The aforesaid algorithms are performed in software on an integrated TSK3000A processor core (IP cores used) and implemented in hardware targeting FPGA (Xilinx Spartan 3AN). The developed platform is validated by generating real-time ECG beats using MIT-BIH arrhythmia database. The performance of the proposed implementation is evaluated in terms of sensitivity, specificity, positive predictivity and accuracy which is 97.72%, 99.09%, 96.46% and 97.96% respectively. The proposed hardware/software implementation has yielded improved results in comparison to the existing methodologies implemented on the FPGA platforms.