深度学习在医学影像中的应用

杨丽洋; 文戈

doi:10.12122/j.issn.1674-4500.2020.02.01

深度学习在医学影像中的应用

doi: 10.12122/j.issn.1674-4500.2020.02.01

杨丽洋,
文戈^,

南方医科大学南方医院影像中心，广东广州 510515

基金项目: 国家自然科学基金(81801682, 81701674);广东省自然科学基金(2020A1515010469);广东省医学科学技术研究基金(C2019077)

详细信息

作者简介:
杨丽洋，硕士，E-mail: 1958084765@qq.com

通讯作者:
文戈，主任医师，博士生导师，E-mail: 1113470826@qq.com

计量
- 文章访问数: 1898
- HTML全文浏览量: 702
- PDF下载量: 88
- 被引次数: 0
出版历程
- 收稿日期: 2020-05-09
- 刊出日期: 2020-04-15

Application of deep learning in medical imaging research

Liyang YANG,
Ge WEN^,

Medical Imaging Center, Nanfang Hospital, Southern Medical University, Guangzhou 510515, China

Funds: Supported by National Natural Science Foundation of China (81801682, 81701674)

摘要

摘要: 医学影像是放射科医生做出医学诊断的重要依据。但随着医学影像技术的快速发展, 逐渐增多的影像图像和复杂的图像信息对医生的工作产生了巨大的挑战。而深度学习是人工智能研究中最热门的领域, 在处理大数据和提取有效信息方面具有优势, 因此逐渐成为分析医学影像方面的首选方法。本文阐述了深度学习的概念, 并简要总结深度学习在医学影像中的常见模型, 包括卷积神经网络、循环神经网络、深度置信网络和自动编码器。卷积神经网络的基本结构是卷积层、池化层和全连接层; 循环神经网络由输入层、隐藏层和输出层组成; 深度置信网络的基础是玻尔兹曼机; 自动编码器包含编码层、隐藏层和解码层。通过对CT肺结节和MRI脑部疾病的分类, 阐明目前深度学习在疾病自动分类上准确性较高; 通过分割左心室、椎旁肌肉和肝脏的结构, 可见深度学习方法在医学图像分割上与人为分割具有一致性; 深度学习在肺结节和乳腺癌疾病的检测上已相对成熟。但目前为止, 仍存在标注的样本量少和过拟合的问题, 希望通过共享图像数据库来解决此问题。总之, 深度学习在医学影像中具有广阔前景, 且对临床医生的工作具有重大意义。
- 深度学习 /
- 医学影像 /
- 神经网络
Abstract: Medical imaging is an important basis for radiologists to make medical diagnosis. However, with the rapid development of medical imaging technology, considerable and complicated medical imaging information brings great challenges to radiologists. Deep learning, as the most popular research field in artificial intelligence, has advantages of processing big data and extracting effective information. Therefore, deep learning has gradually become the top choice for researchers to analyze medical images. Firstly, this paper explains the concept of deep learning. It briefly summarizes the common implementation models of deep learning, including convolutional neural networks, recurrent neural networks, deep belief networks, and automatic encoders. The basic structure of a convolutional neural network is a convolutional layer, a pooling layer, and a fully connected layer. Recurrent neural network is composed of input layer, hidden layer and output layer. The foundation of the deep belief network is the Boltzmann machine. The autoencoder includes coding layer, hidden layer and decoding layer. Then, the classification of CT pulmonary nodules and MRI brain diseases shows that deep learning has high accuracy in automatic disease classification. Deep learning is consistent with artificial segmentation in medical image segmentation by segmenting the left ventricle, paravertebral muscles, and liver. Deep learning is relatively mature in the detection of lung nodules and breast cancer. Finally, the problems of small labeled sample and overfitting are proposed, and we hope to solve this problem by sharing image data. The application of deep learning in medical imaging has bright prospects and great significance on the clinical doctors' work.
- deep learning /
- medical imaging /
- neural networks

HTML全文

参考文献(33)

[1]	Heidenreich A, Desgrandschamps F, Terrier F.Modern approach of diagnosis and management of acute flank pain:review of all imaging modalities[J].Eur Urol, 2002, 41(4):351-62. doi: 10.1016-S0302-2838(02)00064-7/
[2]	Gao J, Jiang Q, Zhou B, et al.Convolutional neural networks for computer-aided detection or diagnosis in medical image analysis:an overview[J].Math Biosci Eng, 2019, 16(6):6536-61. doi: 10.3934/mbe.2019326
[3]	Bonavita I, Rafael-Palou X, Ceresa M, et al.Integration of convolutional neural networks for pulmonary nodule malignancy assessment in a lung cancer classification pipeline[J].Comput Methods Programs Biomed, 2020, 185:105172-8. doi: 10.1016/j.cmpb.2019.105172
[4]	Erickson BJ, Korfiatis P, Akkus Z, et al.Machine learning for medical imaging[J].RadioGraphics, 2017, 37(2):505-15. doi: 10.1148/rg.2017160130
[5]	LeCun Y, Bengio Y, Hinton GE.Deep learning[J].Nature, 2015, 521(7553):436-44. doi: 10.1038/nature14539
[6]	刘飞, 张俊然, 杨豪.基于深度学习的医学图像识别研究进展[J].中国生物医学工程学报, 2018, 37(1):86-94. doi: 10.3969/j.issn.0258-8021.2018.01.012
[7]	夏黎明, 沈坚, 张荣国, 等.深度学习技术在医学影像领域的应用[J].协和医学杂志, 2018, 9(1):10-4. doi: 10.3969/j.issn.1674-9081.2018.01.003
[8]	Litjens G, Kooi T, Bejnordi BE, et al.A survey on deep learning in medical image analysis[J].Med ImageAnal, 2017, 42:60-88. http://d.old.wanfangdata.com.cn/Periodical/gjzdhyjszz-e201806001
[9]	Hubel DH, Wiesel TN.Receptive fields, binocular interaction and functional architecture in the cat's visual cortex[J].J Physiol, 1962, 160(1):106-54. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=10.1113/jphysiol.1962.sp006837
[10]	Fukushima K.Neocognitron:a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position [J].Biol Cybernetics, 1980, 36(4):193-202. doi: 10.1007/BF00344251
[11]	窦瑞欣.深度学习算法在医学影像学中的应用及研究进展[J].中国医学计算机成像杂志, 2018, 24(5):369-72. doi: 10.3969/j.issn.1006-5741.2018.05.001
[12]	Yamashita R, Nishio M, Do RKG, et al.Convolutional neural networks:an overview and application in radiology[J].Insights Imaging, 2018, 9(4):611-29. doi: 10.1007/s13244-018-0639-9
[13]	Maier A, Syben C, Lasser T, et al.A gentle introduction to deep learning in medical image processing[J].Z Med Phys, 2019, 29(2): 86-101. http://cn.bing.com/academic/profile?id=f5082a8741ab9d0fd3329d4072c23a29&encoded=0&v=paper_preview&mkt=zh-cn
[14]	Hinton GE, Salakhutdinov R.Reducing the dimensionality of data with neural networks[J].Science, 2006, 313(5786):504-7. doi: 10.1126/science.1127647
[15]	Saba LC, Biswas M, Kuppili V, et al.The present and future of deep learning in radiology[J].Eur J Radiol, 2019, 114:14-24. doi: 10.1016/j.ejrad.2019.02.038
[16]	Ma JC, Song Y, Tian X, et al.Survey on deep learning for pulmonary medical imaging[J].Front Med, 2019, 120(7):815-26. http://cn.bing.com/academic/profile?id=313b9a0679339ee44395997e0011aaad&encoded=0&v=paper_preview&mkt=zh-cn
[17]	Li L, Liu Z, Huang H, et al.Evaluating the performance of a deep learning-based computer-aided diagnosis (DL-CAD) system for detecting and characterizing lung nodules:Comparison with the performance of double reading by radiologists[J].Thorac Cancer, 2018, 10(2):183-92. http://cn.bing.com/academic/profile?id=52b35c7d855e96cb10bca33ace2f3007&encoded=0&v=paper_preview&mkt=zh-cn
[18]	Zhang S, Han FF, Liang ZR, et al.An investigation of CNN models for differentiating malignant from benign lesions using small pathologically proven datasets[J].Comput Med Imaging Graph, 2019, 77:101645-52. doi: 10.1016/j.compmedimag.2019.101645
[19]	Guan WJ, Ni ZY, Hu Y, et al.Clinical characteristics of coronavirus disease 2019 in China[J].N Engl J Med, 2020.doi:10.1056/ NEJMoa2002032.
[20]	Gao F, Yoon H, Wu T, et al.Afeature transfer enabled multi-task deep learning model on medical imaging[J].Expert Syst Appl, 2020, 143: 112957. doi: 10.1016/j.eswa.2019.112957
[21]	Talo M, Baloglu UB, Yıldırım Ö, et al.Application of deep transfer learning for automated brain abnormality classification using MR images[J].Cogn Syst Res, 2019, 54:176-88. doi: 10.1016/j.cogsys.2018.12.007
[22]	Talo M, Yildirim O, Baloglu UB, et al.Convolutional neural networks for multi-class brain disease detection using MRI images[J]. Comput Med Imaging Graph, 2019, 78:101673-9. doi: 10.1016/j.compmedimag.2019.101673
[23]	Ronneberger O, Fischer P, Brox T.U-net: convolutional networks for biomedical image segmentation[EB/OL].[2015-05-18].https://arxiv.org/abs/1505.04597.
[24]	Lin TY, Dollár P, Girshick R, et al.Feature pyramid networks for object detection[C].2017 IEEE conference on computer vision and pattern recognition (CVPR), 2017: 936-44.
[25]	Gibson E, Li WQ, Sudre C, et al.NiftyNet:a deep-learning platform for medical imaging[J].Comput Methods Programs Biomed, 2018, 158:113-22. doi: 10.1016/j.cmpb.2018.01.025
[26]	Moradi S, Oghli MG, Alizadehasl A, et al.MFP-Unet:a novel deep learning based approach for left ventricle segmentation in echocardiography[J].Phys Med, 2019, 67:58-69. doi: 10.1016/j.ejmp.2019.10.001
[27]	Li HX, Luo HB, Liu YP.Paraspinal muscle segmentation based on deep neural network[J].Sensors, 2019, 19(12):2650-8. doi: 10.3390/s19122650
[28]	Budak U, Guo YH, Tanyildizi E, et al.Cascaded deep convolutional encoder-decoder neural networks for efficient liver tumor segmentation[J].Med Hypotheses, 2020, 134:109431-40. doi: 10.1016/j.mehy.2019.109431
[29]	Ye WJ, Gu W, Guo XJ, et al.Detection of pulmonary ground-glass opacity based on deep learning computer artificial intelligence[J]. Biomed Eng Online, 2019, 18(1):6-14. doi: 10.1186/s12938-019-0627-4
[30]	Zhu HB, Zhao H, Song CH, et al.MR-forest:a deep decision framework for false positive reduction in pulmonary nodule detection[J].IEEE J Biomed Health Inform, 2019, 39(12):1-8. http://d.old.wanfangdata.com.cn/Periodical/txxb201603005
[31]	Ribli D, Horváth A, Unger Z, et al.Detecting and classifying lesions in mammograms with Deep Learning[J].Sci Rep, 2018, 8(1):4165- 71. http://cn.bing.com/academic/profile?id=62f1d2c2f11dad6dcfc9d8ce8affaf06&encoded=0&v=paper_preview&mkt=zh-cn
[32]	McBee MP, Awan OA, Colucci AT, et al.Deep learning in radiology [J].Acad Radiol, 2018, 25(11):1472-80. doi: 10.1016/j.acra.2018.02.018
[33]	Loh BCS, Then PHH.Deep learning for cardiac computer-aided diagnosis:benefits, issues & solutions[J].mHealth, 2017, 10(3):45- 52. http://cn.bing.com/academic/profile?id=f8a92d83d080d67e681b0341e4e63d53&encoded=0&v=paper_preview&mkt=zh-cn