Epicardial and endocardial extracellular potentials imaging has significant implications for diagnosing cardiac diseases. In this paper, we propose a novel noninvasive epicardial and endocardial potentials (EEP) reconstruction method based on low rank and non-local total variation regularization (LRNLTV). The low rank constraint retains the spatiotemporal correlation of EEP and the nonlocal TV constraint make use of the non-local similarity. The constrained optimization problem is solved by augmented Lagrangian multiplier (ALM) method. The simulated PVC data, intervention premature data and clinical PVC data while both EEP map and activation map together with the correlation coefficient (CC), structural similarity (SSIM) and locating error are used to evaluate the proposed method. We also compare LRNLTV method with TV method (L1-norm based) and low rank method, and the proposed framework achieves better results both in reconstruction accuracy and the boundary.
CITATION STYLE
Mu, L., & Liu, H. (2019). Noninvasive Epicardial and Endocardial Extracellular Potentials Imaging with Low-Rank and Non-local Total Variation Regularization. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11902 LNCS, pp. 447–458). Springer. https://doi.org/10.1007/978-3-030-34110-7_37
Mendeley helps you to discover research relevant for your work.