基于改进块贝叶斯学习的多通道心电信号重构

为提高可穿戴心电监护系统的重构精度，本文提出了一种结合多测向量模型的块稀疏贝叶斯学习心电信号重构算法，并在算法的求解过程中使用快速边缘似然最大化算法。对MIT-BIH心律失常数据库、MIT-BIH噪声测试数据库和PTB诊断数据库中心电信号的实验表明，相比于其它传统的压缩感知重构算法，该算法具有重构精度高、运行时间短的优势；相比于基于单测向量模型的块稀疏贝叶斯算法，该算法的重构精度提高了35%，重构速度提高至原来的8倍；在重构含噪声心电信号的情况下，该算法获得比其他重构算法更好的重构效果。因此，本算法在可穿戴心电监护系统中具有良好的应用前景。

Multi-channel ECG reconstruction based on improved Block Sparse Bayesian Learning

In order to improve the reconstruction accuracy of the wearable ECG monitoring system, a CS algorithm based on the Block Sparse Bayesian Learning framework was proposed. It is based on a combination of the block sparse model and multiple measurement vector model. Fast Marginalized Likelihood Maximization is used in the solution process. Experiments on ECG of the MIT-BIH arrhythmia database, The MIT-BIH noise stress test database and PTB diagnostic ECG database show that the proposed algorithm has higher reconstruction accuracy and shorter running time. Compared with the Block Sparse Bayesian Learning based on the single-measure vector model, the reconstruction accuracy of the proposed algorithm is increased by 35%, and the reconstruction speed is increased to 8 times of the original. In the case of reconstructing ECG with noise, it can obtain a better reconstruction effect than other reconstruction algorithms. Therefore, the proposed algorithm has good application prospects in wearable ECG monitoring systems.