基于隐马尔科夫模型的胎盘植入产前诊断方法

发布时间：2018-07-01 21:30

  本文选题：胎盘植入 + 特征提取　； 参考：《模式识别与人工智能》2017年04期

【摘要】：胎盘植入是产科严重的并发症之一,作为金标准的产后病理检验存在的滞后性和局限性问题,文中将病史和彩超数据等产前多特征关联作为观测显状态序列,将产后病理诊断作为隐状态,构建基于隐马尔科夫模型的胎盘植入产前诊断方法.采用Gini方法提取患病关联因素的特征集合,通过转化特征集合构建隐马尔科夫模型,结合Baum-Welch和Viterbi算法计算求解,通过显隐状态关系,实现胎盘植入产前诊断.实验表明,文中方法具有较好的准确率、特异度和灵敏度.
[Abstract]:Placenta accreta is one of the serious complications in obstetrics. As a golden standard of postpartum pathological examination, there are lag and limitation problems. The method of prenatal diagnosis of placenta accreta based on Hidden Markov Model was constructed by using postpartum pathological diagnosis as a hidden state. The Gini method is used to extract the feature sets of the associated factors of the disease, the hidden Markov model is constructed by transforming the feature sets, and the Baum-Welch and Viterbi algorithms are used to calculate and solve the problems. The Prenatal diagnosis of placenta accreta is realized through the explicit hidden state relationship. Experiments show that the proposed method has good accuracy, specificity and sensitivity.
【作者单位】： 福州大学数学与计算机科学学院;福建省妇幼保健院;
【基金】：福建省引导性重点项目(No.2016Y0060,2014Y0005)资助~~
【分类号】：R714.2;TP391.41
，

本文编号：2089215