Implementation of novel ECG compression algorithm using template matching

In this paper, a template based ECG compression algorithm for mobile device application is proposed and implemented. First of all, R-peak is detected in the ECG signal. Then, from the point of R-peak, we extract the information of QRS complex, PQ, and ST section to create a template for each of the interval. Template code is saved for the use of comparing real time input ECG. R-peak is detected in the pre-processing step from a normalized signal in 0 and 1 basis, and adaptive threshold is applied for good R-peak detection accuracy. Template for PQ interval, QRS complex, ST interval is created. Only template code that closely matched with the input signal is saved. The determining factor for `closely matched´ is based on the calculation of MSE value after comparing the template with each of the intervals (PQ interval, QRS complex and ST interval) with real time input ECG signal. The performance of algorithm is evaluated by using the Massachusetts Institute of Technology-Boston´s Beth Israel Hospital Normal Sinus Rhythm Database and different sampling record. The highest compression ratio is 15.2:1. The accuracy of the R-peak detection is 100%.