Research On Signal Processing And Diagnostic Algorithm Of Pathological Electrocardiogram

Signal processing and recognition of electrocardiogram are always animportant issue in the field of modern biomedical detection technology.However, the ECG signal is non-stationary, weak, random and diverse, whichmakes it difficult to be processed especially for a pathological heart. With thedevelopment of computer science, it is available to rapidly calculate complexsteps of modern signal processing methods, which have replaced the position oftraditional ones of ECG detection. In addition, existing portable medicalequipment has been trying to recognize ECG characteristics and diagnose heartdiseases helping for doctors.In order to adapt to diversity of different pathological electrocardiogramsignals, a new method has been developed to improve in diagnosing heartdiseases intelligently by computer in this thesis.First of all, after analyzing the noise of ECG signal, we put forward amethod of thresholding shrinkage of wavelet transformation. Thereby, weestablish a scientific Filtering Experiment Platform under LabVIEWcircumstances and proved the possibility through simulation experiment ofvarious signals in the MIT-BIH Database.Next, combination the advantages of Multilayer extraction algorithmreconstruction and the maximum value of module, we extracted R-wave withdynamical threshold algorithm. We took advantage of the time consistency offormer signals and reconstructed ones, putting forward using Window SweepMethod to extract the Q, S, P and T-wave respectively.Thirdly, we summarize the characteristics of cardiovascular diseasesrelated to the ECG signals, and describe them in the language of numbers andcomputers. By means of the Five Array Pattern Matching Method andpathological feature database, we did simulation diagnose experiments with datain MIT-BIH Database to test the feasibility of this algorithm we designed.We designed the software for digital diagnosis and filtering experimentplatform. And after extracting data from different situation in the experiments, this paper has analyzed and proved the feasibility and reliability of this kind ofnew filtering methods, and then verified the algorithm of all-features extractionthrough simulation of the moving-tested ECG and shaking-tested ECG signals.