| Abstract | | The purpose to participate in the PhysioNet Challenge 2011 was to test our recent computational techniques that might help in assessing of the misplaced ECG electrodes. ECG recordings from the Challenge data sets were analyzed in three steps. First, general characteristics of the recorded signals were determined. Second, cluster analysis was applied to obtain templates that helped to estimate the quality of individual beats, and to isolate the background signal, used for signal drift determination. Finally, equivalent dipoles were determined and used to reconstruct the template. Since its precision depends on the exact location of the lead locations, we speculated that by varying the lead locations to minimize error (optimization) the actual lead location could be detected. We found that the method is reliable to detect correctly placed leads, and less reliable in recognizing different swaps. For the PhysioNet/CinC Challenge 2011 we obtained a score of 0.818.
|