基于通道注意力与空洞卷积的胸片肺气肿检测算法

兰州理工大学学报 ›› 2022, Vol. 48 ›› Issue (2): 81-89.

• 自动化技术与计算机技术 • 上一篇下一篇

基于通道注意力与空洞卷积的胸片肺气肿检测算法

李策^*, 许大有, 靳山岗, 高伟哲, 陈晓雷

兰州理工大学电气工程与信息工程学院, 甘肃兰州 730050

收稿日期:2020-11-23 出版日期:2022-04-28 发布日期:2022-05-07
通讯作者: 李策(1974-),男,辽宁营口人,博士,教授,博导.Email:christ.cli@foxmail.com
基金资助:
国家自然科学基金(61866022,61967012)

Chest X-ray emphysema detection algorithm based on channel attention and dilated convolution

LI Ce, XU Da-you, JIN Shan-gang, GAO Wei-zhe, CHEN Xiao-lei

College of Electrical and Information Engineering, Lanzhou Univ. of Tech., Lanzhou 730050, China

Received:2020-11-23 Online:2022-04-28 Published:2022-05-07

摘要/Abstract

摘要： 胸片中的肺气肿检测算法在临床辅助诊断中具有重要研究意义.针对已有算法缺乏特征通道筛选能力,特征图感受野较小易受局部组织噪声干扰,以及难易样本不均衡等问题,提出了一种基于通道注意力与空洞卷积的胸片肺气肿检测算法EDACD.首先,利用通道注意力模块构建了具有通道选择能力的特征提取网络SE-ResNet及特征金子塔网络SE-FPN.同时,使用空洞卷积代替了SE-ResNet中部分普通卷积,通过增大特征图感受野,提高所提特征鲁棒性.最后,采用焦点损失作为分类损失函数使网络专注困难样本训练.此外,训练过程中采用限制对比度自适应直方图均衡化算法对胸片进行预处理,进一步突出了肺气肿特征.通过数据扩充、聚类标签优化先验框参数的方式,克服肺气肿标注数据稀缺及传统先验框设置不符合肺气肿尺度等问题.在ChestX-Det10和ChestX-Det14数据集中的主、客观实验表明,所提算法较对比算法具有更好的肺气肿病灶检测能力.

关键词: 肺气肿检测, 通道注意力, 空洞卷积, 焦点损失

Abstract: Emphysema detection algorithm in chest X-ray plays an important role in clinical aided diagnosis. In order to solve the problems in existing algorithms, such as lacking feature channel screening ability, small receptive field of feature map, easy to be disturbed by local tissue noise, sample imbalance, EDACD (a chest X-ray emphysema detection algorithm based on channel attention and dilated convolution) is proposed in this paper. Firstly, the feature extraction network SE-ResNet and the feature pyramid network SE-FPN with channel selection ability are constructed by using channel attention module. At the same time, the dilated convolution is used to replace some common convolutions in SE-ResNet, and the robustness of the features is improved by increasing its receptive field. Finally, the focus loss is used as the classification loss function to make the network focus on training difficult samples. In addition, in the training process, the limited contrast adaptive histogram equalization algorithm is used to preprocess the image, which further highlights the characteristics of emphysema. Through data expansion and clustering labels to optimize the anchor parameters, so as to overcome the scarcity of labeled emphysema data and the inappropriate traditional anchors setting of emphysema. The subjective and objective experiment results in Chestx-Det10 and Chestx-Det14 datasets show that the proposed algorithm has better detection ability than the contrast algorithms.

Key words: emphysema detection, channel attention, dilated convolution, focal loss

中图分类号:

TP393

李策, 许大有, 靳山岗, 高伟哲, 陈晓雷. 基于通道注意力与空洞卷积的胸片肺气肿检测算法[J]. 兰州理工大学学报, 2022, 48(2): 81-89.

LI Ce, XU Da-you, JIN Shan-gang, GAO Wei-zhe, CHEN Xiao-lei. Chest X-ray emphysema detection algorithm based on channel attention and dilated convolution[J]. Journal of Lanzhou University of Technology, 2022, 48(2): 81-89.

参考文献

[1] 施俊,汪琳琳,王珊珊,等.深度学习在医学影像中的应用综述 [J].中国图象图形学报,2020,25(10):1953-1981.
[2] CAO L,LI W J,FENG Q J.Automatic detection and diagnosis of lung nodules on CT images based on LDA and SVM [J].Journal of Southern Medical University,2011,31(2):324-328.
[3] REN J.ANN vs.SVM:Which one performs better in classification of MCCs in mammogram imaging [J].Knowledge-Based Systems,2012,26:144-153.
[4] WANG X,PENG Y,LU L,et al.Chestx-ray8:Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2017:2097-2106.
[5] YAO L,POBLENZ E,DAGUNTS D,et al.Learning to diagnose from scratch by exploiting dependencies among labels [J/OL].[2020-10-15].https://arxiv.org/pdf/1710.10501.pdf.
[6] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition [J].Computer Science,2014,1(20):332-342.
[7] HE K M,ZHANG X Y,REN S Q,et al.Deep residual learning for image recognition [C/OL].[2020-10-15].https://arxiv.org/pdf/1512.03385.pdf.
[8] HU J,SHEN L,SUN G.Squeeze-and-excitation networks [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2018:7132-7141.
[9] LIN T Y,GOYAL P,GIRSHICK R,et al.Focal loss for dense object detection [C]//Proceedings of the IEEE International Conference on Computer Vision.[S.l.]:IEEE,2017:2980-2988.
[10] PANG J,CHEN K,SHI J,et al.Libra R-CNN:Towards balanced learning for object detection [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2019:821-830.
[11] 李策,刘昊,陈晓雷,等.基于多感知特征计算的沙尘图像增强算法 [J].兰州理工大学学报,2018,44(4):90-95.
[12] REDMON J,FARHADI A.YOLO9000:better,faster,stronger [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2017:7263-7271.
[13] REN S Q,HE K M,GIRSHICK R,et al.Faster R-CNN:Towards real-time object detection with region proposal networks [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
[14] LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single shot multibox detector [C/OL].[2020-10-15].https://arxiv.org/abs/1512.02325.
[15] ZHAO Q,SHENG T,WANG Y,et al.M2det:A single-shot object detector based on multi-level feature pyramid network [C/OL].[2020-10-15].https:arxiv.org/abs/1811.04533.
[16] LIU J,LIAN J,YU Y.ChestX-det10:chest X-ray dataset on detection of thoracic abnormalities[J/OL].[2020-10-15].http://arxiv.org/abs/2006.10550.

基于通道注意力与空洞卷积的胸片肺气肿检测算法

Chest X-ray emphysema detection algorithm based on channel attention and dilated convolution

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics

本文评价