利用超声图像进行胎盘功能无损自动分级的两种方法

Two Methods of Non-invasive Automatic Classification for the Placental Function Using the Ultrasound Image

根据B型超声得到的不同发育阶段的胎盘图像 ,对胎盘功能进行无损自动分级 .首先从图像中提取出用以胎盘分级的各种特征参数 ,然后结合医学专家的分级判断 ,利用模糊分类法和数量化理论 ,分别建立起自动分级的规则 ,并通过临床试用比较 2种分级方法的优缺点 .结果表明 :这种胎盘功能的无损分级方法可以有效地用于医学临床 .

An non-invasive automatic classification method is proposed based on the B model ultrasound images of the placenta of various gestational weeks. First of all several characteristic parameters are extracted from the images. Then together with the grading results of the doctor, the rules of automatic classification are established by two methods, the fuzzy classification and the quantification theory. Finally the performance of these two methods is compared with the clinical application. The results show that the non-invasive automatic classification method of the placental function can be feasibly used in clinic.