基于卷积神经网络的医学CT图像自动分割方法  

Automatic Segmentation Method of Medical CT Image Based on Convolutional Neural Network

作　　者：胡扬 HU Yang(Information Center,Beijing Tongren Hospital Affiliated to Capital Medical University,Beijing 100730 China)

机构地区：[1]首都医科大学附属北京同仁医院信息中心,北京100730

出　　处：《自动化技术与应用》2024年第3期70-73,共4页Techniques of Automation and Applications

摘　　要：针对医学CT图像自动分割方法对三维分层医学CT图像分割精准度较差问题,设计基于卷积神经网络的医学CT图像自动分割方法。对采集到的原始医学CT图像进行去噪处理,为后续的图像分割提供训练集。设定卷积神经网络的激活函数及网络参数,构建医学CT图像自动分割卷积神经网络。计算医学CT图像在卷积过程中损失函数,获取图像最优分割参数,根据此参数完成分割过程。实验结果表明:此方法无论在图像的整体分割还是细节分割方面,均可获取精度较高的分割结果。Aiming at the problem that the automatic segmentation method of medical CT images has poor segmentation accuracy for 3D layered medical CT images,an automatic segmentation method of medical CT images based on convolutional neural network is de-signed.The acquired original medical CT images are denoised to provide a training set for subsequent image segmentation.The activation function and network parameters of the convolutional neural network are set,and the convolutional neural network for automatic segmentation of medical CT images is constructed.It calculates the loss function of the medical CT image in the convolution process,obtains the optimal segmentation parameter of the image,and completes the segmentation process according to this parameter.The experimental results show that this method can obtain high-precision segmentation results in both the overall segmentation and the detail segmentation of the image.

关 键 词：卷积神经网络 图像分割 统计迭代重建 堆叠降噪自编码网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]