Cerebral arterial thrombosis segmentation system based on three-dimensional convolution

According to the invention, the cerebral arterial thrombosis segmentation difficulty is high; the cerebral arterial thrombosis forms are various; large differences between samples, enabling three-dimensional information, in order to solve the problems that a conventional method based on divergence measurement and the like is low in precision, the invention provides a system based on a 3D convolutional neural network. The system comprises the steps: firstly carrying out the data enhancement and normalization of a training sample, then training a segmentation network with expansion convolution,and carrying out the adjustment on a network structure till an optimal structure network is obtained. Improving segmentation accuracy. 本发明针对缺血性脑卒中分割难度大，缺血性脑卒中形态多样，样本之间差异性大，拥有三维信息，且传统基于散度测度等方法的精度不高的问题，提出一种基于3D卷积神经网络的系统，该系统包括：首先对训练样本进行数据增强，归一化处理，然后训练带有膨胀卷积的分割网络，再在网络结构上调整，直到得到的结构最优网络。提高分割精度。