[1]郭艺,杜秋晨,吴朦朦,等.基于轻量级神经网络的新冠肺炎CT新型识别技术[J].中国医学物理学杂志,2022,39(10):1263-1269.[doi:DOI:10.3969/j.issn.1005-202X.2022.10.014]
　GUO Yi,DU Qiuchen,WU Mengmeng,et al.COVID-19 recognition technology based on lightweight neural network[J].Chinese Journal of Medical Physics,2022,39(10):1263-1269.[doi:DOI:10.3969/j.issn.1005-202X.2022.10.014]
点击复制

基于轻量级神经网络的新冠肺炎CT新型识别技术()
分享到：

《中国医学物理学杂志》[ISSN:1005-202X/CN:44-1351/R]

卷:
39卷
期数:
2022年第10期
页码:
1263-1269
栏目:
医学影像物理
出版日期:
2022-11-02

文章信息/Info

Title:
COVID-19 recognition technology based on lightweight neural network
文章编号:
1005-202X（2022）10-1263-07
作者:
郭艺1; 杜秋晨2; 吴朦朦3; 马鹏涛1; 李冠华1
1.火箭军特色医学中心麻醉科， 北京 100088； 2.北京航空航天大学电子信息工程学院， 北京 100191； 3.火箭军特色医学中心影像科， 北京 100088
Author(s):
GUO Yi1;  DU Qiuchen2;  WU Mengmeng3;  MA Pengtao1;  LI Guanhua1
1. Department of Anesthesiology, PLA Rocket Force Characteristic Medical Center, Beijing 100088, China 2.School of Electronic Information Engineering, Beihang University, Beijing 100191, China 3.Department of Imaging,PLA Rocket Force Characteristic Medical Center, Beijing 100088, China
关键词:
新冠肺炎; 电子计算机断层扫描; 轻量级网络; 识别技术; GhostNet
Keywords:
Keywords: corona virus disease 2019 computed tomography lightweight network recognition technology GhostNet
分类号:
R318；TP391
DOI:
DOI:10.3969/j.issn.1005-202X.2022.10.014
文献标志码:
A
摘要:
目的：为了满足临床新冠肺炎检测的实际需求，提出一种基于轻量级人工神经网络的新冠肺炎CT新型识别技术。方法：首先，选取目前公开的所有新冠肺炎CT图像数据集，经过图像亮度规范化和数据集清洗后作为训练数据，通过大样本提高深度学习的泛化能力；其次，采用GhostNet轻量级网络简化网络参数，使深度学习模型能够在医用计算机上运行，提高新冠肺炎CT诊断的效率；再次，在网络输入中加入肺部区域分割图像，进一步提高新冠肺炎CT诊断的准确性；最后，提出加权交叉熵损失函数减小漏诊率。结果：在本研究构建的数据集上进行测试，所提出方法的精确率、召回率、准确率和F1值分别为83%、96%、90%和88%，且在医用计算机上耗时为236 ms。结论：本研究提出方法的效率和准确性均优于其他对比算法，能较好地适应新冠肺炎诊断的需求。
Abstract:
Abstract: Objective To propose a novel corona virus disease 2019 (COVID-19) recognition technology based on lightweight neural network for meeting the actual needs of COVID-19 detection in clinic. Methods All public COVID-19 CT image data set were selected and taken as training data after gray-level normalization and data cleaning. The generalization ability of deep learning was improved by large sample. Then the lightweight network GhostNet was adopted to simplify the network parameters, so that the deep learning model could run on medical computer and improved the efficiency of COVID-19 diagnosis based on CT. Subsequently, the diagnostic accuracy was further improve by adding lung image segmentation to network input. Finally, a weighted cross-entropy loss function was used to reduce the rate of missed diagnosis. Results The proposed method was tested on data set constructed in this study. The precision, recall rate, accuracy and F1 value of the proposed method were 83%, 96%, 90% and 88% respectively, and it took 236 ms to complete COVID-19 recognition on medical computer. Conclusion The proposed method is superior to other algorithms in efficiency and accuracy, and it can better meet the needs of COVID-19 diagnosis.

相似文献/References:

[1]蔡爱楠,随力,王君. CT/MRI双模态造影剂的制备及研究进展[J].中国医学物理学杂志,2018,35(2):219.[doi:DOI:10.3969/j.issn.1005-202X.2018.02.020]
　CAI Ainan,SUI Li,WANG Jun. Preparation and research progress of CT/MRI bimodal contrast agent[J].Chinese Journal of Medical Physics,2018,35(10):219.[doi:DOI:10.3969/j.issn.1005-202X.2018.02.020]
[2]窦益腾,朱新进,夏俊,等. 64排螺旋CT对结节性甲状腺肿的诊断价值[J].中国医学物理学杂志,2018,35(3):303.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.011]
　DOU Yiteng,ZHU Xinjin,XIA Jun,et al. Diagnostic value of 64-slice spiral CT for nodular goiter[J].Chinese Journal of Medical Physics,2018,35(10):303.[doi:DOI:10.3969/j.issn.1005-202X.2018.03.011]
[3]何重阳,文明.CT在外突性子宫肌瘤与卵巢性索间质肿瘤的鉴别诊断价值[J].中国医学物理学杂志,2020,37(5):589.[doi:10.3969/j.issn.1005-202X.2020.05.011]
　HE Chongyang,WEN Ming.Diagnostic value of CT in outprojecting uterine leiomyoma and ovarian sex cord-stromal tumor[J].Chinese Journal of Medical Physics,2020,37(10):589.[doi:10.3969/j.issn.1005-202X.2020.05.011]
[4]黄勃,王志.后踝撞击综合征MRI和CT诊断价值分析[J].中国医学物理学杂志,2020,37(6):730.[doi:DOI:10.3969/j.issn.1005-202X.2020.06.014]
　HUANG Bo,WANG Zhi.Diagnostic values of MRI and CT in posterior ankle impingement syndrome[J].Chinese Journal of Medical Physics,2020,37(10):730.[doi:DOI:10.3969/j.issn.1005-202X.2020.06.014]
[5]刘胜,范承武.CT诊断重症肺炎支原体肺炎的价值[J].中国医学物理学杂志,2021,38(7):842.[doi:DOI:10.3969/j.issn.1005-202X.2021.07.010]
　LIU Sheng,FAN Chengwu.Value of CT in diagnosing severe Mycoplasma pneumoniae pneumonia[J].Chinese Journal of Medical Physics,2021,38(10):842.[doi:DOI:10.3969/j.issn.1005-202X.2021.07.010]
[6]李娜,曾珉,刘俊宏.脑利钠肽水平联合肺动脉CT评估老年心力衰竭患者预后的价值[J].中国医学物理学杂志,2021,38(9):1119.[doi:10.3969/j.issn.1005-202X.2021.09.013]
　LI Na,ZENG Min,LIU Junhong.Value of brain natriuretic peptide level combined with pulmonary artery CT in evaluating theprognosis of elderly patients with heart failure[J].Chinese Journal of Medical Physics,2021,38(10):1119.[doi:10.3969/j.issn.1005-202X.2021.09.013]
[7]周美君,朱晟.高频超声与CT在良恶性浅表淋巴结鉴别诊断中的应用价值[J].中国医学物理学杂志,2022,39(4):464.[doi:DOI:10.3969/j.issn.1005-202X.2022.04.013]
　ZHOU Meijun,ZHU Sheng.Value of high-frequency ultrasound and CT in the differential diagnosis of benign and malignant superficial lymph node lesions[J].Chinese Journal of Medical Physics,2022,39(10):464.[doi:DOI:10.3969/j.issn.1005-202X.2022.04.013]
[8]翁羽洁,李忠贤,姬宇程,等.基于改进阈值的VGG网络的新冠肺炎CT图像自动诊断算法[J].中国医学物理学杂志,2022,39(6):731.[doi:DOI:10.3969/j.issn.1005-202X.2022.06.013]
　WENG Yujie,LI Zhongxian,JI Yucheng,et al.Automatic diagnosis algorithm for COVID-19 CT images using improved threshold-based VGG network[J].Chinese Journal of Medical Physics,2022,39(10):731.[doi:DOI:10.3969/j.issn.1005-202X.2022.06.013]
[9]乔文俊,周芳,刘泉芬,等.深度学习图像重建算法对改善直肠CT图像质量的临床应用价值[J].中国医学物理学杂志,2024,41(8):975.[doi:DOI:10.3969/j.issn.1005-202X.2024.08.008]
　QIAO Wenjun,ZHOU Fang,et al.Improving rectal CT image quality with a deep learning image reconstruction algorithm[J].Chinese Journal of Medical Physics,2024,41(10):975.[doi:DOI:10.3969/j.issn.1005-202X.2024.08.008]

备注/Memo

备注/Memo:
【收稿日期】2022-03-18 【作者简介】郭艺，住院医师，研究方向：医学图像处理，E-mail: bonita183@126.com
更新日期/Last Update: 2022-10-27