Header menu link for other important links
X
Robust and accurate personalised reconstruction of standard 12-lead system from Frank vectorcardiographic system
S. Maheshwari, , P.E. Puddu, E.B. Mazomenos, M. Schiariti, K. Maharatna
Published in Taylor and Francis Ltd.
2016
Volume: 4
   
Issue: 3-4
Pages: 183 - 192
Abstract
In this article, we have proposed a robust and accurate method for the reconstruction of standard 12-lead (S12) system from Frank vectorcardiographic (FV) system using personalised transformation (PT) matrices targeting personalised remote health monitoring applications. FV system is used in the 3D visualisation of heart and diagnosis and prognosis of many cardiologic disorders including myocardial infarction, Brugada syndrome etc. However, cardiologists are accustomed to S12 system pertaining to its decades-old usage and widespread acceptability and hence, it is generally used as primary ECG acquisition system and state-of-the-art inverse Dower transform (DT) and affine transform (AT) are used to obtain FV system from S12 system. Here, we propose the acquisition of FV system and use PT to reconstruct S12 system from FV system. PhysioNet's PTB database after wavelet-based preprocessing to remove baseline wandering and noise has been used in this investigation. The personalised coefficients have been obtained using least-squares fit method and heart-vector projection theory and evaluation metrics used are R2 statistics, correlation and regression coefficients. Our proposed PT methodology has outperformed AT and DT by mean R2 values of 16.36% and 26.53%, respectively. For the practical application of the proposed system, we have investigated into the reusability of personalised coefficients which has been shown to outperform state-of-the-art AT and DT. © 2014 Informa UK Limited, trading as Taylor & Francis Group.
About the journal
JournalData powered by TypesetComputer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization
PublisherData powered by TypesetTaylor and Francis Ltd.
ISSN21681163