一种改进的基于压缩感知的心电压缩算法

An Improved Compressed Sensing-Based ECG Compression Algorithm

摘要:

为了解决远程动态心电记录仪数据量过大的问题，并且克服基于压缩感知的压缩算法压缩比有限的问题，提出压缩感知压缩与移位差分位压缩结合的心电数据压缩算法，移位差分位压缩算法是无损压缩算法，在不影响压缩感知重构精度的前提下进一步提高压缩比。经实验证明，该方法将压缩感知原有4倍的压缩比最高提高到11倍，最小为4.81倍，压缩端的计算复杂度为O(N),满足远程动态心电记录仪的需求。

Abstract:

In order to solve the problem of huge data in remote holter recording, ensure the accuracy of ECG data reconstruction and reduce compression power consumption, an algorithm that combines compressive sensing compression with shift differential bit compression is proposed. Shift difference bit compression is a lossless compression algorithm, which improves the compression ratio without affecting the reconstruction accuracy. Block sparse Bayesian model is used to reconstruct ECG signals, the reconstruction accuracy meets the needs of physicians for diagnosis. Experiments show that the maximum compression ratio obtained by this method is 11, the minimum value is 4.81, and the computational complexity of the compression end is O(N). This satisfies the requirements of remote dynamic ECG recorders.