脑电记录中数据压缩的研究

STUDY ON EEG DATA COMPRESSION

近年内脑电压缩的研究很少见,而心电压缩技术有很大的进展。脑电在频率范围和波形特征上和心电有相近之处。故作者从心电压缩算法中选择合适的、较好的扇形算法作为脑电压缩研究的起点,实验结果表明其效果不佳。分析该算法的缺点后,提出一种新的算法:四选一算法。对脑电数据压缩得到比较满意的结果。在压缩比和百分均方根误差方面,明显地优越。处理速度快,能对16道脑电的数据进行实时处理。本文叙述新算法的原理和实验结果。

<ABSTRACT> In the field of ECG data compression, Blanchard and Barr presented results showing that the fan metheod was clearly superior to other popular and commercial algorithm. We tried to use the fan method to EEG data compression, but the results were not as good as expected. Drawbacks of the fan method were analysed and a new algorithm- 'select one from four' was proposed. For EEG data compression, the new algorithm is clearly su- perior in terms of data reduction rate and percent RMS difference. Exe- cution speed is so fast that 16 channels EEG data compression can run in real time.