A novel QRS complex detection on ECG with motion artifact during exercise

We present a novel QRS complex detection scheme from ECG with motion artifact. The algorithm relies on subspace learning and template matching. QRS complex detection during exercise is a challenging problem because multiple artifacts affect the ECG measurement. Motion artifact is considered to be the main disturbance added to the measurement during exercise. To deal with the problem, we train a dictionary to represent motion artifact using information from a tri-axis accelerometer, and then remove the artifact contribution from noisy ECG measurements. We select the GCC-PHAT filter for efficient QRS detection on the denoised ECG measurements. We show that the proposed algorithm has appreciably higher motion artifact reduction capability and lower computational complexity than competing algorithms. It is therefore a preferred alternative for implementation in mobile health monitoring systems.