Compression of ECG Signal Using Neural Network Predictor and Huffman Coding

Abstract

Medical signals and images need special treatment especially when the data become bigger and bigger. One of the treatment that will be considered in this experiment is about the data compression. ECG (Electro Cardio Graph) signals will create very big data when the signals were collected in a long period of time. Several methods can be used to compress the ECG data. In this experiment we used neural network to predict the incoming data and huffman coding to minimize the codes. The ECG data was collected from MIT-BIH arrhythmia database. The experiment gave low compression rasio when the predicted data was very close to the incoming data.