A CS framework of data reduction is proposed for multichannel ECG (MECG) signals in eigenspace.

PCA is used to exploit the spatial correlation across the channels resulting into sparse eigenspace signals.

Using the compressed sensing (CS) approach, the significant eigenspace signals are gone through further dimensionality reduction.

OMP is used for the CS recovery by exploiting the eigenspace/other domain sparsity of the PCA transformed MECG signals.

The approach leads to higher compression efficiency, which makes it useful for resource-constrained MECG telemonitoring applications.