基于功率谱特征的心电数据小波压缩方法

Compression of ECG Data based on Power Spectral Characteristics and Wavelet Transform

利用心电功率谱特征,探索心电数据压缩新方法。用小波分解心电信号为高频与低频分量,对低频分量继续分解达到要求的级数,对高频分量则根据其所在频段的能量,对临床诊断的价值加以取舍。对MIT生理信号数据库心电数据的压缩与还原分析表明,该方法平衡了压缩比与还原精度之间的矛盾,既具有较高的压缩比,又具有较高的还原精度,而且对信号的适应性也明显增强。另外,该压缩方法还具有一定的去噪作用。说明结合心电功率谱特征与小波变换方法压缩心电有其优势。

To propose a new approach for ECG data compression using power spectral characteristics of ECG and wavelet transform. Firstly, ECG signal was decomposed into low-frequency and high-frequency part by using wavelet transform. Then decomposition was continued for low-frequency part until required times was reached. According to the value in clinical diagnosis of the different energy bands in the high-frequency part, reservation or reset of these signal bands were performed. The test results showed that this approach proposed had better performance than traditional DCT compression algorithm, beth compression rate and reconstruction precision were improved. In addition, the new algorithm had enhanced adaptability to almost all ECG signals and played some roles in noise cancellation. It proves that the proposed approach is a potential tool in ECG data compression.