Title :
BSeg++: A modified Blind Segmentation Method for Ballistocardiogram Cycle Extraction
Author :
Akhbardeh, A. ; Kaminska, Bozena ; Tavakolian, Kouhyar
Author_Institution :
Tampere Univ. of Technol., Tampere
Abstract :
This paper presents a method to extract cardiac cycles and H-I-J components of Ballistocardiogram (BCG). The new improved algorithm BSeg++ permits on the segmentation of BCG signal and extraction of its basic complexes H-I-J without Electrocardiogram (ECG) synchronization. The BSeg++ is based on two previously developed methods described in [1, 2, 3] for extracting BCG cycles without using a reference ECG signal. Those methods suffered from extract redundant BCG cycles because of motion artifacts or BCG fluctuations. In this study, we modified the blind segmentation algorithm and solved its problems. We also added another feature to detect H-I-J complexes of BCG. Also, this new algorithm can be used to extract cardiac cycles and R-S-T components of ECG. The data analysis has been performed on the subjects tested at Simon Fraser University. Initial tests of BCG and ECG from twenty subjects indicate that the method extracted BCG (ECG) cycles and its components with a negligible error in the presence of motion artifacts, BCG fluctuations, latency and non-linear disturbance.
Keywords :
electrocardiography; medical signal processing; BSeg++; ballistocardiogram cycle extraction; cardiac cycles; electrocardiogram; modified blind segmentation method; motion artifacts; Band pass filters; Blood; Data mining; Electrocardiography; Fluctuations; Force measurement; Force sensors; Heart; Signal processing; Testing; Algorithms; Artifacts; Ballistocardiography; Data Compression; Data Interpretation, Statistical; Electrocardiography; Equipment Design; Heart Diseases; Humans; Models, Statistical; Motion; Reproducibility of Results; Signal Processing, Computer-Assisted; Software;
Conference_Titel :
Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th Annual International Conference of the IEEE
Conference_Location :
Lyon
Print_ISBN :
978-1-4244-0787-3
DOI :
10.1109/IEMBS.2007.4352686