Design Of ECG Automatic Analysis And Diagnosis System

In the biomedical engineering field, the study of ECG waveform characteristics has been a hotspot,at the same time, the realization of automatic ECG analysis and diagnosis capabilities will greatly promote China's medical industry.In this paper, a ECG signal analysis and diagnosis system has designed for the practical application,which was also a software engineering tool for ECG automatic detection and extraction. Algorithms in engineering researched around three main directions,the "ECG signal preprocessing", "ECG detection and extraction", "Design and implementation of ECG analysis software". Firstly, a variety of methods used in pretreatment and extraction of ECG feature have been introduced, and compared these methods to explain the advantages and disadvantages. Followed by detailed analysis of these methods used in ECG signal preprocessing of the VC project. Then this paper analyzed the methods of extraction characteristics of the ECG waveform and feature point, and proposed two extraction methods of QRS wave feature points:First one based on the variable differential threshold,and it's a high-efficiency method of detection of QRS wave group characteristics and software implementation; The second one was based on multi-modal morphological derivative (Multi-Modal Morphological Derivatives, MMD) transform to detect QRS wave group's initial point and ending point. In the detection of T wave's initial point and ending point,this paper used the 'wing' function to do T wave's transformation, because 'wing'function's local amplification,T wave can be more precise positioned.Finally,researched on ECG automatic classification methods and determined appropriate branching logic methods for engineering application.Then,used computer language to express and classify the clinical experiences, and designed reasonable procedures to realize ECG automatic diagnosis function.Database used in this paper was internationally recognized MIT-BIH Arrhythmia Database, and the algorithms simulation tool was matlab 7.0. Algorithms' simulation of 48 Arrhythmia signals verified the accuracy and reliability of methods in this paper. Finally, this paper used the computer language C++and object-oriented method to design software, and used VisualC++6.0 on windows platform to write software. This software included the display of ECG waveform, the analysis of characteristic parameters, and diagnosis reports function.