胎心宫缩图参数分析和胎儿状态评估方法的研究

发布时间：2018-03-24 10:40

本文选题：胎心宫缩图　切入点：胎心率基线　出处：《暨南大学》2015年硕士论文

【摘要】：近年来,我国的出生缺陷率持续上升,计划生育、优生优育国策面临严峻挑战。实施胎儿监护,有助于及时发现胎儿异常,从而降低出生缺陷率和婴儿死亡率。胎心宫缩图(CTG)是目前最常见的胎儿监护方法。通过分析CTG信号,可以对胎儿的健康状况进行评估。然而,很多医护工作者判读CTG的水平不足,容易误判而做出不当的临床决策。CTG计算机辅助分析的出现,在一定程度上缓解了这个问题。但是,由于CTG的复杂性以及参数定义不够严谨,计算机分析仍存在很多问题,参数识别的准确率有待提高。为此,本文研究CTG关键参数的提取方法,以提高参数识别的准确率和效率;在此基础上,本文进一步研究胎儿状态评估方法,提出一种更为有效的评估方法。本文的主要研究内容如下:(1)胎心率基线估计算法的研究。针对现有算法准确性或效率不高的问题,本文提出了一种结合胎动的胎心率基线估计算法。在该算法中,利用胎动确定胎心率曲线中加速段的位置,将其去除后进行基线估计,再对基线进行修正。将该算法与两种现有算法对比,结果表明:在分析准确性方面,本文算法明显胜过参照算法一,稍微优于参照算法二;而在计算效率方面,本文算法与参照算法一相差不大,但远高于参照算法二。(2)宫缩曲线特征提取算法的研究。本文研究了宫缩基线和宫缩波识别方法,并提出了相应的新方法。与一种现有的宫缩基线估计算法相比,本文算法可得到更平滑的基线,并且能实现断点检测,从而使最终结果更合理。本文结合图像膨胀原理和宫缩波形态学分析,提出了一种宫缩波识别算法。通过与一种现有算法作对比,结果显示:本文算法的整体识别效果好于参照算法,且错判率、漏判率均低于参照算法。同时,本文还设计了一种宫缩状态实时识别方法,可实现四种状态的实时判别。通过比较本算法的实时分析结果与医生的事后分析结果,发现两者吻合度很高。而且,该算法可满足实时分析对计算效率的要求。(3)胎儿状态评估方法的研究。针对目前计算机辅助分析系统直接套用CTG分类标准而导致胎儿状态分析结果不准的问题,本文根据文献调研和实验结果对CTG分类标准进行修正,并结合模糊集合的思想,计算各个CTG参数对不同状态的隶属度,并使用欧几里得距离衡量CTG信号与三种标准状态间的差距,从而实现胎儿状态的分析。实验结果表明,与直接应用分类标准的方法相比,本文方法可识别出更多的正常类信号,且该方法的特异度和阳性预测值远高于参照方法,总体准确率也明显高于参照方法。
[Abstract]:In recent years, the rate of birth defects in China has been rising, and the national policy of family planning and eugenics is facing severe challenges. The implementation of fetal monitoring is conducive to the timely detection of fetal abnormalities. Thus reducing the rate of birth defects and infant mortality. CTG is the most common method of fetal monitoring at present. By analyzing the CTG signal, the health of the fetus can be evaluated. However, Many health care workers interpret the level of CTG is insufficient, easy to misjudge and make improper clinical decision. CTG computer-aided analysis to some extent alleviates this problem. However, because of the complexity of CTG and parameter definition is not strict. There are still many problems in computer analysis, and the accuracy of parameter identification needs to be improved. In order to improve the accuracy and efficiency of parameter identification, this paper studies the extraction method of key parameters of CTG. In this paper, we further study the fetal state assessment method and propose a more effective evaluation method. The main contents of this paper are as follows: 1) the research on the baseline estimation algorithm of fetal heart rate. In this paper, an algorithm of fetal heart rate baseline estimation combined with fetal movement is proposed. In this algorithm, the position of accelerated segment in the fetal heart rate curve is determined by fetal movement, and the baseline estimation is carried out after removing it. The results show that the proposed algorithm is superior to the reference algorithm 1, slightly better than the reference algorithm 2 in terms of accuracy of analysis, and the efficiency of the algorithm is higher than that of the reference algorithm 2. The difference between this algorithm and reference algorithm one is not different, but much higher than that of reference algorithm 2. 2) the method of feature extraction of uterine contraction curve is studied in this paper. The method of identifying uterine contraction baseline and uterine contraction wave is studied in this paper. Compared with one of the existing base-line estimation algorithms, the proposed algorithm can obtain a smoother baseline and can detect breakpoints. Combining the principle of image expansion and morphological analysis of uterine constriction wave, this paper proposes an algorithm for recognition of uterine constriction wave, which is compared with one of the existing algorithms. The results show that the overall recognition effect of this algorithm is better than that of the reference algorithm, and the error rate and missing rate are lower than that of the reference algorithm. At the same time, a real-time recognition method of uterine contraction is designed. The real-time discriminant of four states can be realized. By comparing the real time analysis results of this algorithm with the results of doctors' post event analysis, it is found that the degree of agreement between the two methods is very high. The algorithm can meet the requirements of real-time analysis for computing efficiency. It can be used to study the method of fetal state evaluation. Aiming at the problem that the current computer-aided analysis system directly applies CTG classification standard, the result of fetal state analysis is inaccurate. In this paper, the classification standard of CTG is modified according to the literature investigation and experimental results, and the membership degree of each CTG parameter to different states is calculated by combining the idea of fuzzy set. The Euclidean distance is used to measure the difference between the CTG signal and the three standard states, so that the fetal state can be analyzed. The experimental results show that the proposed method can recognize more normal signals than the direct classification method. The specificity and positive predictive value of the method were much higher than that of the reference method, and the overall accuracy was significantly higher than that of the reference method.
【学位授予单位】：暨南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R714.5

【参考文献】