Efficient ECG modeling using polynomial functions

Abstract

An efficient ECG modeling algorithm is presented in this paper. The model is based on fitting polynomial functions to real ECG. This algorithm describes a segmentation of a heartbeat and fitting an appropriate polynomial function to the segments. The model performances are evaluated in terms of PRD, preserve of ST-T segment clinical information and the execution time. When comparing this model with the existing one, the PRD improvements can be seen, especially in those signals with high morphological diversity heartbeats. Moreover, the computing time is significantly reduced. Using appropriate model features, ST-T analysis achieves an average accuracy of more than 94%. The obtained data shows that this model is applicable to other ECG processing like: heartbeat analysis, compression and filtering.