基于活动轮廓模型和影像组学的脑卒中图像预测方法

Stroke image prediction method based on active contour model and radiomics

针对脑卒中患者的神经功能缺损程度预测问题,提出活动轮廓模型和影像组学的方法。首先提出基于模糊速度函数的活动轮廓模型方法对梗死病灶区分割。然后提取形态学、一阶统计和纹理特征,接着用LASSO方法来特征选择。最后采用距离加权KNN方法构建分类预测模型。结果表明,提出的分割算法获得的结果接近金标准,分类算法有较好的预测性能。

A method of active contour model and radiomics is proposed to predict the degree of neurologic impairment of stroke patients.An active contour model based on fuzzy velocity function is proposed to segment infarct area.The morphology,first⁃order statistics and textural features are extracted,and then the least absolute shrinkage and selection operator(LASSO)method is used for the feature selection.The distance⁃weighted k nearest neighbor(KNN)algorithm is used to construct the classification prediction model.The results show that the result obtained by the proposed segmentation algorithm is close to the gold standard,and the classification algorithm had better predictive performance.