基于级联网络的肝脏肿瘤CT图像分割  

CT Image Segmentation of Liver Tumor Based on Cascade Network

作　　者：莫亚霓 陈晓婕 张本鑫 MO Yani;CHEN Xiaojie;ZHANG Benxin(School of Mathematics and Computing Science,Guilin University of Electronic Technology,Guilin 541004,China;School of Computer and Information Security,Guilin University of Electronic Technology,Guilin 541004,China;School of Electronic Engineering and Automation,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学数学与计算科学学院,广西桂林541004 [2]桂林电子科技大学计算机与信息安全学院,广西桂林541004 [3]桂林电子科技大学电子工程与自动化学院,广西桂林541004

出　　处：《电视技术》2024年第1期38-41,共4页Video Engineering

基　　金：国家级大学生创新训练项目(202210595041)。

摘　　要：肝脏肿瘤计算机断层扫描(Computed Tomography,CT)图像分割是肝癌诊断与治疗过程的重要环节。近年来,基于U型结构的卷积神经网络在该分割任务取得了巨大的成功,但仍存在一些挑战,如肿瘤边界分割不精确、小肿瘤难以检测等。为提高肝脏肿瘤的分割精度,提出一种级联网络MCPUNet用于肝脏肿瘤分割任务。MCPUNet引入MDB(MDconv Block)和MP(Mixing Pooling)以获取上下文信息,MDB通过混合深度可分离卷积和坐标注意力机制提取图像的多尺度特征,MP用于下采样减少图像尺寸。实验证明,与原始的U-Net模型相比,该模型在肝脏肿瘤分割任务上的交并比(Intersection over Union,IoU)、准确度和召回率指标分别提高3.8%、2.5%和2.0%,为肝癌诊断和治疗提供了可靠依据。Liver tumor Computed Tomography(CT)image segmentation is a crucial step in the diagnosis and treatment of liver cancer.In recent years,convolutional neural networks based on the U-shaped structure have achieved significant success in this segmentation task,but there are still some challenges,such as inaccurate tumor boundary segmentation and difficulty in detecting small tumors.In order to improve the segmentation accuracy of liver tumors,this paper proposes a cascade network called MCPUNet for liver tumor segmentation tasks.MCPUNet introduces MDB(MDconv Block)and MP(Mixing Pooling)to obtain contextual information,MDB extracts multi-scale features of the image by mixing depth-wise separable convolutions and coordinate attention mechanisms,while MP is used for downsampling to reduce the image size.Experimental results show that compared to the original U-Net model,this model has improved Intersection over Union(IoU),precision,and recall metrics by 3.8%,2.5%,and 2.0%,respectively,in liver tumor segmentation tasks,providing a reliable basis for the diagnosis and treatment of liver cancer.

关 键 词：肝脏肿瘤分割 混合深度可分离卷积 级联网络 多尺度 注意力机制 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]