基于解绕Fourier分解的远程心电图实时分析研究

开展远程心电图实时分析研究，对心血管疾病患者的及时诊断具有重要的理论价值和现实意义。然而现有研究存在心电图信号压缩效率低、时间-频率特征提取方法精确度差、分类器单一等不足。本课题针对远程心电图信号压缩、时间-频率特征提取、分类器的设计三方面展开研究。拟分析解绕Fourier分解方法的快速收敛性和计算复杂度，通过存储心电图信号的分解参数来构建新类型的压缩模型，以获得高效实时压缩；提出解绕Fourier分解的时间-频率表示理论，给出心电图信号的时间-频率特征提取新方法，使获取的时间-频率特征更适合于分类应用；从多分类器融合以弥补单一分类器不足的角度，探索两分类器恰当融合的新方法，提出协同表示分类和最近质心分类相融合的新分类器，以实现心电图信号的高效识别。课题有望在心电图信号实时压缩与识别等方面取得理论和技术突破，为远程心电图监护系统的进一步开发提供理论依据和技术支持。

Carrying out real-time analysis of remote electrocardiogram (ECG) has important theoretical value and practical significance for the timely diagnosis of patients with cardiovascular disease. However, the existing studies have shortcomings such as low compression efficiency of ECG signals, poor accuracy of time-frequency feature extraction methods, and single classifier algorithms. This project focuses on three aspects of remote ECG signal compression, time-frequency feature extraction, and design of classifiers. The project will analyze the fast convergence and computational complexity of the unwinding Fourier decomposition (UFD) method, and then build a new type of compression model by storing the decomposition parameters of the ECG signals to obtain efficient real-time compression. The project will propose the time-frequency representation theory of UFD. Based on the proposed theory, a new method for extracting time-frequency features of ECG signals will be presented to make the derived time-frequency features more suitable for classification. From the perspective of multi-classifier fusion to compensate for the shortcomings of a single classifier, the project will explore new methods for the proper fusion of two classifiers. A new classifier that combines collaborative representation classification and nearest centroid will be developed to achieve effective and efficient recognition of ECG signals. This project is expected to achieve theoretical and technical breakthroughs in real-time compression and recognition of ECG signals, and provide theoretical and technical support for the further development of remote ECG monitoring systems.