Application of two non-linear methods for the localization of artifacts on trend data [medical]
Lotfi Senhadji, Nikos Nikolaidis, Jaakko Ojaniemi · 2002
This paper presents the application of two non-linear methods for the localization of transient artifacts on trend data. The first method is based on the matching pursuit method. The second is a multi-dimensional technique which makes use of a particular neural network, namely the marginal median linear vector quantizer. The application concerns trend data issued from monitored pulmonary artery pressure.