Paper Info
Title
Spatial enhancement of ECG using diagnostic similarity score based lead selective multi-scale linear model.
Abstract
In this work, a new patient-specific approach to enhance the spatial resolution of ECG is proposed and evaluated. The proposed model transforms a three-lead ECG into a standard twelve-lead ECG thereby enhancing its spatial resolution. The three leads used for prediction are obtained from the standard twelve-lead ECG. The proposed model takes advantage of the improved inter-lead correlation in wavelet domain. Since the model is patient-specific, it also selects the optimal predictor leads for a given patient using a lead selection algorithm. The lead selection algorithm is based on a new diagnostic similarity score which computes the diagnostic closeness between the original and the spatially enhanced leads. Standard closeness measures are used to assess the performance of the model. The similarity in diagnostic information between the original and the spatially enhanced leads are evaluated using various diagnostic measures. Repeatability and diagnosability are performed to quantify the applicability of the model. A comparison of the proposed model is performed with existing models that transform a subset of standard twelve-lead ECG into the standard twelve-lead ECG. From the analysis of the results, it is evident that the proposed model preserves diagnostic information better compared to other models.
Year
DOI
Venue
2017
10.1016/j.compbiomed.2017.04.002
Computers in Biology and Medicine
Keywords
Field
DocType
00–01,99-00
Data mining,Pattern recognition,Computer science,Closeness,Linear model,Selection algorithm,Correlation,Artificial intelligence,Image resolution,Wavelet,Repeatability
Journal
Volume
ISSN
Citations 
85
0010-4825
1
PageRank 
References 
Authors
0.39
13
2
Name
Order
Citations
PageRank
Nallikuzhy, Jiss J.141.13
S. Dandapat226128.51