Ambulatory Holter ECG Individual Events Delineation via Segmentation of a Wavelet-Based Information-Optimized 1-D Feature

The aim of this study is to develop and describe a new ambulatory holter electrocardiogram (ECG) events detection-delineation algorithm via segmentation of an information-optimized decision statistic. After implementation of appropriate pre-processing, a uniform length sliding window is applied to the pre-processed trend and in each slide, some geometrical features of the excerpted segment are calculated to construct a newly proposed Discriminant Analyzed Geometric Index (DAGI), by application of a nonlinear orthonormal projection. Then the a-level Neyman-Pears on classifier is implemented to detect and delineate QRS complexes. The presented method was applied to several databases and the average values of sensitivity and positive predictivity, Se = 99.96% and P+ = 99.96%, were obtained for the detection of QRS complexes, with an average maximum delineation error of 5.7 msec, 3.8 msec and 6.1 msec for P-wave, QRS complex and T-wave, respectively. Also the method was applied to DAY general hospital high resolution holter data (more than 1,500,000 beats, including Bundle Branch Blocks- BBB, Premature Ventricular Complex-PVC, and Premature Atrial Complex-PAC) and average values of Se=99.98% and P+ = 99.97% were obtained for QRS detection. High accuracy in a widespread SNR, high robustness and processing speed (1^6,000 samples/sec) are important merits of the proposed algorithm.