Real-time classification of heartbeats using least square acceleration filter for ambulatory monitoring

This paper describes a novel algorithm for real-time heartbeat classification based on the least square acceleration filter (LSAF) and the linear discriminant function (LDF). QRS complex was detected by LSAF with 30th order and the features were extracted. The performance of the proposed algorithm was evaluated using the MIT-BIH Arrhythmia database. The results showed the detection rate of 99% of the QRS complex and classification rate of 97% of various heartbeats such as normal beat, premature ventricular contraction, atrial premature contraction, left bundle branch block beat, and right bundle branch block beat. The algorithm was implemented on DSP chip in real time. We also compared the performance of the floating point simulation with that of fixed point simulation. There was no difference in the performance of the classification between the floating and fixed point simulation. But the computational cost of fixed point simulation was remarkably reduced than that of floating point simulation. It showed the possibilities that the proposed algorithm can be used effectively in the ambulatory monitoring.