过参数卷积与CBAM融合的胸腔积液肿瘤细胞团块分割网络  

Over-parametric convolution and attention mechanism-fused pleural effusion tumor cell clump segmentation network

作　　者：陈思卓 赵萌[1,2] 石凡 黄薇 Chen Sizhuo;Zhao Meng;Shi Fan;Huang Wei(Engineering Research Center of Learning-Based Intelligent System,Ministry of Education,Tianjin 300384,China;School of Computer Science and Engineering,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]学习型智能系统教育部工程研究中心,天津300384 [2]天津理工大学计算机科学与工程学院,天津300384

出　　处：《中国图象图形学报》2023年第10期3243-3254,共12页Journal of Image and Graphics

基　　金：国家自然科学基金项目(62020106004,92048301,61703304);大学生创新创业训练计划项目(202010060038)。

摘　　要：目的胸腔积液肿瘤细胞团块的分割对肺癌的筛查有着积极作用。胸腔积液肿瘤细胞团块显微图像存在细胞聚集、对比度低和边界模糊等问题,现有网络模型进行细胞分割时无法达到较高精度。提出一种基于UNet网络框架,融合过参数卷积与注意力机制的端到端语义分割模型DOCUNet(depthwise over-parameterized CBAM UNet)。方法将UNet网络中的卷积层替换为过参数卷积层。过参数卷积层结合了深度卷积和传统卷积两种卷积,保证网络深度不变的同时,提高模型对图像特征的提取能力。在网络底端的过渡区域,引入结合了通道注意力与空间注意力机制的注意力模块CBAM(convolutional block attention module),对编码器提取的特征权重进行再分配,增强模型的分割能力。结果在包含117幅显微图像的胸腔积液肿瘤细胞团块数据集上进行5折交叉实验。平均IoU(inter⁃section over union)、Dice系数、精确率、召回率和豪斯多夫距离分别为0.8580、0.9204、0.9282、0.9203和18.17。并且与UNet等多种已存在的分割网络模型进行对比,IoU、Dice系数和精确率、召回率相较于UNet提高了2.80%、1.65%、1.47%和1.36%,豪斯多夫距离下降了41.16%。通过消融实验与类激活热力图,证明加入CBAM注意力机制与过参数卷积后能够提高网络分割精度,并能使网络更加专注于细胞的内部特征。结论本文提出的DOCUNet将过参数卷积和注意力机制与UNet相融合,实现了胸水肿瘤细胞团块的有效分割。经过对比实验证明所提方法提高了细胞分割的精度。Objective Lung cancer-related early detection and intervention is beneficial for lowering mortality rates.Pleu⁃ral effusion symptoms,and tumor cells and tumor cell masses can be sorted out in relevant to pleural effusion and its meta⁃static contexts.The detection of tumor cells in pleural effusion can be recognized as an emerging screening tool for lung can⁃cer for early-stage intervention.One of the key preprocessing steps is focused on the precise segmentation of tumor cell masses in related to pleural fluid tumor cells.However,due to severe tumor cell masses-between overlapping and adhe⁃sion,unclear cell-to-cell spacing,and unstable staining results of tumor cells in pleural effusion are challenged to be resolved using conventional staining methods,manual micrographs of unstained pleural fluid tumor clumps for cell clump segmentation derived of experienced and well-trained pathologists.But,it still has such problems of inefficiency and the inevitable miss segmentation due to its labor-intensive work.In recent years,computer vision techniques have been devel⁃oping intensively for optimizing the speed and accuracy of image analysis.Traditional methods for segmenting cellular microscopic images are carried out,including thresholding and such algorithms of clustering-based,graph-based,and active contouring.However,these methods are required for image downscaling,and they have the limitations of undevel⁃oped graphical features.Convolutional neural network(CNN)based deep learning can be used to automatically find suit⁃able features for image segmentation tasks nowadays.The UNet is derived from the end-to-end full convolutional network(FCN)structure,and it is widely used in medical image segmentation tasks due to its unique symmetric encoder and decoder network structure to get the segmentation result relevant to location information of the segmented target,in which arbitrary size image input and equal size output image can be yielded for arbitrary size image input and equal size output image.We develop

关 键 词：胸腔积液肿瘤细胞团块 UNet 注意力机制 细胞分割 过参数卷积 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]