Transmisión optimizada de electrocardiogramas para efectos de clasificación

Abstract

This paper presents a design for optimization of transmission process of ECG signal that will be analyzed remotely for classification. The optimization process is based in the reduction of information that is transfered and the reduction of time in the execution of the application that clasifies. The model presented is tested with an implementation of a client/server application. The client application take the ECG, compress it and transfers it to the remote server application that use the same parameters that used by the client application for compression and with them continue the classification process. This is achived used wavelets for the characterization of ECG signal. The compression results are analyzed in terms PRD reliability (root mean square difference) and compression rates. The signal intelligibility achived after the uncompression process is evaluated with the classification algorithm. The compress rate achived is 84.7% and 86% of successful for Bayesian classifier. © Springer-Verlag Berlin Heidelberg 2007.