COVID-19 X-Ray Diagnosis Method Based on Improved Global Contextual Attention

doi:10.3778/j.issn.1002-8331.2207-0279

Abstract

Abstract: In the task of diagnosing novel coronavirus pneumonia, the irrelevant information of chest X-ray（CXR） will affect the model classification decision. Segmentation network is used to extract lung parenchyma first and then classification is an effective way. A two-stage classification model Res-IgSa is proposed. The lung parenchyma of CXR images is extracted first by using the resunet network. The classification network takes ResNet50 as the basic framework, and introduces the improved global context module（WGC） and spatial attention module（CSA）. WGC introduces channel normalization and gating adaptive unit based on the global context to better adjust the relationship between channels. CSA introduces group convolution, and controls the capacity of the model by the number of groups. WGC and CSA can realize the complementarity between characteristic graphs. A large number of experiments are carried out on COVID-19 radio databaseV5 dataset. Compared with the same method used in the original paper, Res-IgSa achieves better classification effect, and the accuracy, precision, recall and F1 value reach 94.154%, 94.157%, 94.154% and 94.139% respectively. Ablation experiments and result visualization further prove the effectiveness of the model.

Key words: corona virus disease 2019（COVID-19）, deep learning, segmentation, classification, attention mechanism, CXR image

摘要： 在新型冠状病毒肺炎诊断任务中，胸部X射线（chest X-ray，CXR）的无关信息会影响模型分类决策，利用分割网络先将肺实质提取出来再进行分类是一种有效的途径。提出一种两阶段的分类模型Res-IgSa，使用ResUNet网络先提取CXR图像的肺实质，分类网络以ResNet50为基本框架，引入改进全局上下文模块（WGC）以及空间注意力模块（CSA）。WGC在全局上下文的基础上引入通道规范化以及门控自适应单元，以更好地调整通道间的关系；CSA引入分组卷积，通过分组数来控制模型的容量，WGC和CSA可以实现特征图间的互补。在COVID-19 Radiography DatabaseV5数据集上进行了大量的实验，与原论文采用相同的方法相比，Res-IgSa实现了更好的分类效果，准确率、精度、召回率以及F1值分别达到了94.154%、94.157%、94.154%以及94.139%，并进行消融实验以及结果可视化进一步证明了该模型的有效性。

关键词: 新型冠状病毒肺炎, 深度学习, 分割, 分类, 注意力机制, CXR图像

JI Xurui, LIU Jing, JI Hui, ZHANG Shuai, CAO Hui. COVID-19 X-Ray Diagnosis Method Based on Improved Global Contextual Attention[J]. Computer Engineering and Applications, 2023, 59(21): 222-230.

吉旭瑞, 刘静, 吉辉, 张帅, 曹慧. 改进全局上下文注意力新冠肺炎X光诊断方法[J]. 计算机工程与应用, 2023, 59(21): 222-230.

References

[1] CHEN N，ZHOU M，DONG X，et al.Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan，China：a descriptive study[J].The Lancet，2020，395（10223）：507-513.
[2] XIE X，ZHONG Z，ZHAO W，et al.Chest CT for typical 2019-nCoV pneumonia：relationship to negative RT-PCR testing[J].Radiology，2020，296（2）：200343.
[3] RAJPURKAR P，IRVIN J，ZHU K，et al.Chexnet：radiologist-level pneumonia detection on chest x-rays with deep learning[J].arXiv.1711.05225，2017.
[4] ZECH J R，BADGELEY M A，LIU M，et al.Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs：a cross-sectional study[J].PLoS Medicine，2018，15（11）：e1002683.
[5] FENG Y，YANG X，QIU D，et al.PCXRNet：pneumonia diagnosis from chest X-ray images using condense attention block and multiconvolution attention block[J].IEEE Journal of Biomedical and Health Informatics，2022，26（4）：1484-1495.
[6] 刘锐，丁辉，尚媛园，等.COVID-19医学影像数据集及研究进展[J].计算机工程与应用，2021，57（22）：15-27.
LIU R，DING H，SHANG Y Y，et al.COVID-19 medical imaging dataset and research progress[J].Computer Engineering and Applications，2021，57（22）：15-27.
[7] SOUSA P E D，CARNEIRO P C，OLIVEIRA M M，et al.COVID-19 classification in X-ray chest images using a new convolutional neural network：CNN-COVID[J].Research on Biomedical Engineering，2021，38：87-97.
[8] LOEY M，SMARANDACHE F，KHALIFA N.A deep transfer learning model with classical data augmentation and CGAN to detect COVID-19 from chest CT radiography digital images[J].Neural Computing and Applications，2020：1-13.
[9] RAJPAL S，KUMAR N，RAJDAL A，et al.COV-ELM classifier：an extreme learning machine based identification of COVID-19 using chest-ray images[J].arXiv：2007. 08637v5，2020.
[10] WANG L，LIN Z Q，WONG A.COVID-Net：a tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images[J].Scientific Reports，2020，10（1）.
[11] MAITY A，NAIR T R，MEHTA S，et al.Automatic lung parenchyma segmentation using a deep convolutional neural network from chest X-rays[J].Biomedical Signal Processing and Control，2022，73：103398.
[12] XU Y，LAM H K，JIA G.MANet：a two-stage deep learning method for classification of COVID-19 from chest X-ray images[J].Neurocomputing，2021，443.
[13] KALAIVANI S，SEETHARAMAN K.A three-stage ensemble boosted convolutional neural network for classification and analysis of COVID-19 chest x-ray images[J].International Journal of Cognitive Computing in Engineering，2022，3（1）：35-45.
[14] KARIM M R，D?HMEN T，COCHEZ M，et al.Deep COVID explainer：explainable COVID-19 diagnosis from chest X-ray images[C]//2020 IEEE International Conference on Bioinformatics and Biomedicine（BIBM），2020：1034-1037.
[15] ZHANG Z，LIU Q，WANG Y.Road extraction by deep residual U-Net[J].IEEE Geoscience and Remote Sensing Letters，2018，15（5）：749-753.
[16] 欧阳柳，贺禧，瞿绍军.全卷积注意力机制神经网络的图像语义分割[J].计算机科学与探索，2022，16（5）：1136-1145.
OUYANG L，HE X，QU S J.Fully convolutional neural network with attention module for semantic segmentation[J].Journal of Frontiers of Computer Science and Technology，2022，16（5）：1136-1145.
[17] RONNEBERGER O，FISCHER P，BROX T.U-Net：convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention，2015：234-241.
[18] WANG X，GIRSHICK R，GUPTA A，et al.Non-local neural networks[C]//2018 IEEE Conference on Computer Vision and Pattern Recognition，2018：7794-7803.
[19] CAO Y，XU J，LIN S，et al.GCNet：non-local networks meet squeeze-excitation networks and beyond[C]//2019 IEEE/CVF International Conference on Computer Vision Workshop（ICCVW），2019：1971-1980.
[20] YANG Z，ZHU L，WU Y，et al.Gated channel transformation for visual recognition[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR），2020：11791-11800.
[21] GUO M H，XU T X，LIU J J，et al.Attention mechanisms in computer vision：a survey[J]Computational Visual Media，2022，8：331-368.
[22] 卓天天，桑庆兵.注意力机制与复合卷积在手写识别中的应用[J].计算机科学与探索，2022，16（4）：888-897.
ZHUO T T，SANG Q B.Application of attention mechanism and composite convolution in handwriting recognition[J].Journal of Frontiers of Computer Science and Technology，2022，16（4）：888-897.
[23] 吴静然，丁恩杰，崔冉，等.采用多尺度注意力机制的旋转机械故障诊断方法[J].西安交通大学学报，2020，54（2）：51-58.
WU J R，DING E J，CUI R，et al.A diagnostic approach for rotating machinery using multi-scale feature attention mechanism[J].Journal of Xi’an Jiaotong University，2020，54（2）：51-58.
[24] IOANNOU Y，ROBERTSON D，CIPOLLA R，et al.Deep roots：improving CNN efficiency with hierarchical filter groups[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2017：5977-5986.
[25] LIU Y，SHAO Z，HOFFMANN N.Global attention mechanism：retain information to enhance channel-spatial Interactions[C]//IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2021.
[26] WOO S，PARK J，LEE J Y，et al.CBAM：convolutional block attention module[C]//European Conference on Computer Vision，2018.
[27] PARK J，WOO S，LEE J Y，et al.BAM：bottleneck attention module[J].arXiv：1807.06514，2018.
[28] CHOWDHURY M E H，RAHMAN T，KHANDAKAR A，et al.Can AI help in screening viral and COVID-19 pneumonia?[J].IEEE Access，2020（8）：132665-132676.
[29] RAHMAN T，KHANDAKAR A，QIBLAWEY Y，et al.Exploring the effect of image enhancement techniques on COVID-19 detection using Chest X-rays images[J].Computers in Biology and Medicine，2021，132.
[30] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2016：770-778.
[31] JIE H，LI S，GANG S，et al.Squeeze-and-excitation networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42（8）：2011-2023.
[32] SELVARAJU R R，COGSWELL M，DAS A，et al.Grad-CAM：visual explanations from deep networks via gradient-based localization[C]//2017 IEEE International Conference on Computer Vision（ICCV），2017：618-626.