Principal component analysis based method for reconstruction of fragments of corrupted or lost signal in multilead data reflecting electrical heart activity and hemodynamics

摘要

Multilead signals reflecting electrical activity of the heart and hemodynamics give comprehensive but usually redundant representation of the processes. Therefore fragments of transient corruption or loss of the data in one or more leads can be restored substituting them by the signals reconstructed using information carried by the other leads. Principal component analysis used for reduction of dimensionality in representation of the physiological processes. It concentrates essential information represented by multilead signal into few principal components which could form a set of basis functions for optimal representation and reconstruction of the original signals. The idea of the method is to calculate these components from normal fragment of the multilead signal and use them for reconstruction of missing lead in the corrupted fragment of it.