基于U-Net网络的肺部CT图像分割算法

Design of smart campus system based on cloud computing English title

针对CT图像存在双肺边界难以准确分割、肺实质周围存在血管和细小空洞时,传统分割方法无法得到良好的分割效果这一问题,文中提出了一种基于U-Net的肺部CT图像自动分割算法。首先对原CT图像进行高斯和拉普拉斯滤波处理,然后将预处理后的图像和原始图像作为输入,分别使用U-Net进行分割。最后将所有分割出来的肺部区域通过线性回归进行融合从而提取出肺部实质区域。定性与定量实验结果表明,文中使用的基于U-Net的分割方法能有效分割肺部实质区域,且该算法相比于传统算法更加出色。

In order to solve the problem that the presence of pulmonary or pleural dense regions in CT image, and traditional image segmentation methods cannot get good segmentation results, this paper proposes an algorithm for automatic segmentation of lung CT image based on U-Net. This mehod firstly process the original CT image by adopting Gauss and Laplasse filter;then the preprocessed image and the original image as input and segmentation by U- Net ;finally all the segmented lung regions were fused to extract pulmonary parenchyma by linear regression. The results of qualitative and quantitative experiments show that the U-Net based segmentation method can effectively segment the lung parenchyma, and the algorithm is superior to the traditional algorithm.