三维多尺度嵌套U结构CT影像肺结节检测  

Detection of pulmonary nodules in three-dimensional multiscale nested U-structure computed tomography images

作　　者：许正玺 张少敏[1,2,3] 支力佳 周涛[1,3] Xu Zhengxi;Zhang Shaomin;Zhi Lijia;Zhou Tao(School of Computer Science and Engineering,North Minzu University,Yinchuan 750021,China;Medical Imaging Center,Ningxia Hui Autonomous Region People’s Hospital,Yinchuan 750000,China;Key Laboratory of Images&Graphics Intelligent Processing of State Ethnic Affairs Commission,North Minzu University,Yinchuan 750021,China)

机构地区：[1]北方民族大学计算机科学与工程学院,银川750021 [2]宁夏回族自治区人民医院医学影像中心,银川750000 [3]北方民族大学图像图形智能处理国家民委重点实验室,银川750021

出　　处：《中国图象图形学报》2022年第3期797-811,共15页Journal of Image and Graphics

基　　金：国家自然科学基金项目(61561002,62062003);宁夏自然科学基金项目(2020AAC03213,2021AAC03198);宁夏自治区重点研发计划项目(2020BEB04022);“图像与智能信息处理创新团队”国家民委创新团队项目(PY1606,PY1905);宁夏医学影像临床研究中心创新平台建设项目(2018DPG05006);北方民族大学一般科研项目(2021XYZJK04)。

摘　　要：目的肺结节检测在低剂量肺部计算机断层扫描(computed tomography,CT)筛查肺癌中具有重要意义。但由于结节大小、形状和密度的变化十分复杂,导致难以在低假阳性率下保证高的灵敏度,这限制了深度学习算法在常规临床实践中的肺结节自动诊断,建立具有良好结节检测性能的深度学习模型仍然是一个挑战。针对此问题,本文提出了一种基于3D ReSidual U(3D RSU)块的嵌套U结构的肺结节检测框架。方法3D RSU块通过混合不同大小的感受场获得多尺度特征来丰富特征信息。而嵌套U结构允许网络获得更大分辨率的特征图,从而具有多层次深度特征,获得丰富的局部和全局信息,增强网络区分前景和背景的能力,进而提高微小结节等非显著性目标的检测性能。结果该框架在公共肺结节(lung nodule analysis 16)挑战数据集上进行了评价。方法能够准确地检测出肺结节,灵敏度达到了97.2%,与基准方法相比,该方法灵敏度提高了2.6%,具有很高的灵敏度和特异性,在0.125、0.25、0.5、1、2、4、8共7个假阳率点的灵敏度平均值为86.4%,尤其是在每扫描0.25个假阳性上,灵敏度达到80.9%,优于基准算法76.9%的结果。结论本文所提出的结节检测模型,由于在低假阳率上有较高的灵敏度,可以使本网络在常规临床实践中为医生辅助诊断提供更加可靠清晰的早期肺癌参考信息。Objective Lung cancer is one of the most common diseases in humans and mainly causes the rising mortality rate.Medical experts believe that the early diagnosis of lung cancer can reduce mortality by screening for lung nodules through computed tomography(CT).Checking a large number of CT images can reduce lung cancer risk.However,CT scan images contain a large volume of information about nodules,and as the number of images increases,accuracy becomes a very challenging task for radiologists.Therefore,an effective computer-aided diagnosis(CAD)system should be designed.However,the high false-positive rate remains a challenging problem.In response,this paper proposes a new lung nodule detection framework based on 3 D ReSidualU-blocks(3 D RSU)module and nested U structure.Method This paper trains an end-to-end lung nodule detection model for one stage.The whole system can shorten the processing time of pulmonary nodule detection without reducing the accuracy of early detection.The 3 D deep convolutional neural network(CNN)proposed in this paper is based on the region proposal network(RPN)of the Faster R-CNN Network.The 3 D CNN can make full use of the spatial information of CT images.The problem of missed detection by the nodule and the large number of false-positive nodules during detection of nodules with small diameter nodules can be solved by dividing the front background information and enhancing the ability to detect non-salient objects in the image.In this paper,the 3 D RSU was designed to capture multi-scale features within the stage.The symmetrical codec structure can be used to learn how to extract and encode multi-scale context information.By using the different layers of 3 D RSU,the network can allow feature maps of any spatial resolution as input elements to extract elements at multiple scales.This process reduces the loss of detail caused by large-scale direct up-sampling.The 3 D RSU was embedded into the network to form a nested u-shaped structure.This structure allows the network to obtain larger resolu

关 键 词：肺结节检测 嵌套结构 特征金字塔网络 非显著性目标检测 10折交叉验证 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]