ECG Data Compression And QRS Wave Recognition

The electrocardiogram(ECG) is an important tool for checking abnormal information about the heart. Data compression and signal recognition of ECG take important role in cardiac computerized analysis. The development of algorithms promotes the technology to make incressant progress in this field.The purpose of this thesis is to recognize several typical QRS wave by using neural network, and to provide a new approach for ECG data compression based on wavelet neural network.Firstly, the wavelet neural network is established to compress ECG data. The result shows that the given compression algorithm has good performance of fast constringency speed and highly compressionradio. The network reserve the vital information about ECG data so that it can help to diagnosing.Secondly, Matlab neural network toolbox is applied to construct training network and to offer fast learning algorithm, several typical QRS waves can be recognized, which proved that this approach conduce to resolve the problem and decrease the difficulty of programming.in addition, in order to make the acquisition of wave data, image processing method and the threshold approach are given to resolve data sampling and detect R wave.