Hybrid time-frequency adaptive filtering for the recovery of ventricular late potentials

Ventricular late potentials are small ECG components that have significant diagnostic importance. Temporal averaging is used to recover these signals from high levels of background noise, but is not the optimum approach. A new form of hybrid time-frequency adaptive filtering is introduced that is based on signal division into a series of frequency bands. This technique is able to process the averaged data further and to reduce significantly the noise levels beyond that achieved by simple averaging. Examples of the processing are shown and its applicability is discussed.