室性早搏的多模板匹配自适应识别算法

Adaptive algorithm for PVC recognition based on multiple template matching

为实现室性早搏的自动判别,提出一种基于多模板匹配的自适应识别算法。采用滤波提高数据信噪比,利用样本数据通过特定的自学习机制建立模板库,设计模板的压缩算法实现待测心搏与模板心搏宽度的对齐,采用模板队列匹配和模板库匹配相结合的匹配策略,设计自适应更新规则实时优化队列与库的结构,利用相关系数检测室早。使用MIT-BIH数据库进行测试,达到99.39%的灵敏度和98.79%的特异度。

An adaptive algorithm for PVC recognition based on multiple template matching is presented. Filtering technology is adapted to improve the SNR of ECG data and a template database is established with training samples through self-learning method. The correlation coefficient is used to detect PVC beats from normal beats, with a compressed method specially designed to adjust the width of the tested heartbeat and the template. A two-stage strategy is proposed based on template queue matching and template database matching and a self-adaptive method is implemented to optimize the two data structures. The performance of this algorithm is evaluated on the MIT-BIH arrhythmia database and the achieved sensitivity for PVC is 99.39% while the specificity is 98.79%.