妊娠期胎盘超声图像分形特征的数量化分析

Quantitative analysis of the fractal parameters gained from the B-mode ultrasound placenta image

应用分数布朗运动模型 (FBM) ,对孕妇不同妊娠期的胎盘B型超声图像进行分形处理 ,提取出用以胎盘分级的分形特征参数 ,并结合临床医学专家的判断 ,利用数量化理论 ,建立起胎盘功能的自动分级模型。通过 10 6例胎盘图像的分析 ,结果表明 :胎盘B型超声图像的分形特征可以有效地表征胎盘的功能状况 ,而数量化理论可以使胎盘功能的无损自动分级成为可能 ,因此有较好的临床应用前景。

Based on Fractional Brown Motion theory, the B mode ultrasound image from the placenta during gestational period is analyzed, and two fractal parameters are obtained. Then together with the grading results judged by medical experts, the rules of automatic classification are established by applying the quantification theory. The application results of 106 ultrasound images of placenta show that the fractal feature extracted from the B mode ultrasound image is a valuable parameter for the placental function, and it is feasible in clinic to evaluate automatically the placental function using the proposed method.