Dynamic Electrocardiograph has provided an essential evidence for diagnosis of cardiovascular disease. It is significant both for telemedici-ne monitoring and clinical diagnosis. To realize real-time dynamic mon-itoring in network environment, real-time dynamic electrocardiogram cognition and automatic diagnosis system must be studied at first.In this paper, based on the study of dynamic electrocardiograph signal processing techniques, multi-resolution and multi-scale charact-eristics of wavelet transform was adopted in the quantitative analysis of electrocardiograph signal preprocessing and extraction of characteristic values. And BP neural network was adopted in the identification of acu-te myocardial ischemia symptoms. The automatic diagnosis system was realized based on Visual C++ 6.0 build environment, using MySQL database and MATLAB.First, Wavelet decomposition and reconstruction method, threshold method and a novel approach combining the two methods were used in the de-noising of original ECG signal. And then an algorithm was desig-ned for waveform recognition and measurement based on the technique of wavelet based break point detection approach. At last, BP neural net-work was used in the identification of dynamic electrocardiograph sign-als. In sum, we realized dynamic electrocardiograph pre-processing, wa-veform recognition, and characteristic value extraction. Simply by read the input ECG signal, disease diagnosis reports can be generated.The study of this paper has established a good foundation for furt-her design and realization of real-time dynamic monitoring in network environment. |