一种自动分割股骨区域的R-U-Net神经网络

R-U-Net Neural Network for Automatic Segmentation of Femur Area

针对目前X线片骨组织分割主要靠医生手工标记耗时且准确率不稳定的问题,本文以股骨为研究对象,提出一种结合深度残差网络和U-Net架构优势的R-U-Net神经网络,并将其应用于股骨区域自动分割.首先对原图像预处理后标注目标区域制作标签图,并对训练集进行数据增强;接着将其输入到R-U-Net神经网络训练调参,保存优化后的网络模型;最后将待分割的测试图像输入到保存的网络模型中得到股骨区域的轮廓,填充连通区域后进行自定义掩码操作得到股骨区域的分割结果.以PhotoShop人工分割结果作为参考,测试结果表明基于R-U-Net的分割效果优于传统方法和U-Net,能够实现批量股骨区域的自动分割,执行效率更高,分割效果更佳.

In order to solve the problem of the time-consuming manual labelling and the unstable accuracy of the X-ray bone tissue,taking the femur as the research object,this paper presents the R-U-Net neural network combining deep residual network and U-Net architecture,and applies it to the automatic segmentation of femoral area. Firstly,after pre-processing of the original dataset,made the label images in the target area,and augmented the training dataset. Then,inputted it to the R-U-Net neural network for training and adjusting parameter,and saved the optimized network model. Finally inputted the test images to be segmented into the saved network model to obtain the outline of the femur area,and performed the mask operation after filling the connected area to obtain the femur area segmentation results. The results of artificial segmentation using PhotoShop as a reference,the test results of R-U-Net are superior to traditional and U-Net based segmentation methods. Realized the automatic segmentation of the mass femur area. It has better execution efficiency and segmentation effect.