Khalfen E Sh, Rybak O K, Nazarov S I
Kardiologiia. 1977 Jul;17(7):40-9.
Because when fed to an on-line computer the ECG information carries a good deal of noise, of prime importance is devising a method for representation and filtration of the useful asignal. It is shown that in order to separate the ueful signal with a protracted direct ECG input from the patient to a computer the use of the proposed modified method of coherent accumulation is advisable. This modification envisages a step-wise solution of the following problems: current diagnosis of arrhythmic contractions exclusion of arrhythmic contractions from the accumulation procedure, automatic search of the normal duration of the RR intervals, equalization of the RRh intervals duration, statistical averaging of the obtained quantum values. The devised modification of the coherent accumulation method, while considerably increasing the noise-immunity of the algorhythm, enables it to effect a direct long-term automatic analysis of the ECG at intensive care units. The identification and measurement of the ECG waves and intervals are done through and automatic selection of the curve elements search zones, finding local maxima and also of the commencement and end of waves by using the method of the "moving window" with an adaptive evaluation of the first signal derivative intensity.
因为当将心电图(ECG)信息输入在线计算机时会携带大量噪声,所以设计一种用于表示和过滤有用信号的方法至关重要。结果表明,为了将患者长时间直接输入计算机的心电图中的有用信号分离出来,建议使用所提出的改进的相干累积方法。这种改进设想逐步解决以下问题:心律失常收缩的当前诊断、从累积过程中排除心律失常收缩、自动搜索RR间期的正常持续时间、使RRh间期持续时间均等化、对获得的量子值进行统计平均。所设计的相干累积方法的改进在显著提高算法抗噪声能力的同时,使其能够在重症监护病房对心电图进行直接长期自动分析。心电图波形和间期的识别与测量是通过自动选择曲线元素搜索区域、找到局部最大值以及使用“移动窗口”方法并对第一信号导数强度进行自适应评估来确定波形的起始和结束来完成的。