HomeArticles2022 No.1Details

Realization of medical image data transfer learning based on MATLAB

Update:Jan. 21, 2022Total Views:1246Total Downloads:1384 DownloadMobile

Author: HUANG Xiaxuan 1, 2 HUANG Tao 2 YUAN Shiqi 1, 2 HE Ningxia 2 WU Wentao 3 LYU Jun 2

Affiliation: 1. Deparment of Neurology, The First Affiliated Hospital of Jinan University, Guangzhou 510630, China 2. Deparment of Clinical Research, The First Affiliated Hospital of Jinan University, Guangzhou 510630, China 3. School of Public Health, Xi'an Jiaotong University, Xi'an 710061, China

Keywords: MATLAB Transfer learning Image classification Pleural effusion

DOI:10.12173/j.issn.1004-5511.202109018

Reference:Huang XX, Huang T, Yuan SQ, He NX, Wu WT, Lyu J. Realization of medical image data transfer learning based on MATLAB[J]. Yixue Xinzhi Zazhi, 2022, 32(1): 33-39. DOI: 10.12173/j.issn.1004-5511.202109018.[Article in Chinese]

  • Abstract
  • Full-text
  • Figures and Tables
  • References
Abstract

Objective  To discuss how to implement the application of medical image data on transfer learning based on MATLAB.

Methods  Taking MIMIC-CXR for example, 500 X-ray images of positive and the same number negative pleural effusion were randomly selected as the total data set, and the MATLAB software transfer learning method based on the ResNet network model was used for multiple training, calculating the AUC value to evaluate the accuracy of the model training for pleural effusion image classification.

Results  The pleural effusion imaging test set and training set used in this study were evenly distributed. Some training models had an accuracy rate of 80%, and the loss rate dropped to below 20%. The highest accuracy rate in training with 250 iterations was up to 100%, which took about 2 minutes and 38 seconds. Based on the prediction results of image data transfer learning obtained in this model training, the highest AUC value was 93.53%.

Conclusion  The transfer learning method of the ResNet network model can realize the effective combination and enhance-ment of model construction and medical imaging data training, and the model has good predictive performance, which provides a certain basis for clinicians in the early diagnosis of pleural effusion.

Full-text
Please download the PDF version to read the full text: download
References

1.陈莹. 迁移学习: 教AI提取抽象知识[N]. 科技日报,2018-1-8(8).

2.刘靖雯, 黄理灿. 基于 Inception-ResNet-V2的乳腺癌病理图像识别研究[J]. 软件导刊, 2020, 19(5): 225-229. [Liu JW, Huang LC. Pathological image recognition of breast cancer based on Inception-ResNet-V2[J]. Software Guide, 2020, 19(5): 225-229.] DOI: 10.11907/rjdk. 192019.

3.何雪英, 韩忠义, 魏本征. 基于深度学习的乳腺癌病理图像自动分类[J]. 计算机工程与应用, 2018, 54(12): 121-125. [He XY, Han ZY, Wei BZ. Breast cancer histopathological image auto-classification using deep learn-ing[J]. Computer Engineering and Applications, 2018, 54(12): 121-125.] DOI: 10.3778/j.issn.1002-8331.1701- 0392. 

4.李莉, 黄韬, 王新宇, 等. 胸腔X射线影像数据库——MIMIC-CXR数据探索[J].中国循证心血管医学杂志, 2021, 13(6): 653-656, 660. [Li L, Huang T, Wang XY, et al. Thoracic X-ray image database-MIMIC-CXR data explora-tion[J]. Chinese Journal of Evidence-Bases Cardiovascular Medicine, 2021, 13(6): 653-656, 660.] DOI: 10.3969/j.issn.1674-4055.2021.06.04.

5.Wu WT, Li YJ, Feng AZ, et al. Data mining in clinical big data: the frequently used databases, steps, and methodological models[J]. Mil Med Res, 2021, 8(1): 44.  DOI: 10.1186/s40779-021-00338-z. 

6.Yang J, Li Y, Liu Q, et al. Brief introduction of medical database and data mining technology in big data era[J]. J Evid Based Med, 2020, 13(1): 57-69. DOI: 10.1111/jebm. 12373.

7.刘栋, 李素, 曹志冬. 深度学习及其在图像物体分类与检测中的应用综述[J]. 计算机科学, 2016, 43(12): 13-23. [Liu D, Li S, Cao ZD. State-of-the-art on deep learning and its application in image object classification and detection[J]. Computer Science, 2016, 43(12): 13-23.] DOI: CNKI:SUN:JSJA.0.2016-12-004.

8.李炳臻, 刘克, 顾佼佼, 等. 卷积神经网络研究综述[J].计算机时代, 2021, (4): 8-12, 17. [Li BZ, Liu K, Gu JJ, et al. Review of the researches on convolutional neural networks[J]. Computer Era, 2021, (4): 8-12, 17.] DOI: 10.16644/j.cnki.cn33-1094/tp.2021.04.003.

9.Alom MZ, Taha TM, Yakopcic C, et al. The history began from AlexNet: a comprehensive survey on deep learning approaches[J/OL]. (2018-09-12) [2021-09-17]. https://arxiv.org/abs/1803.01164.

10.Simonyan KZ. Very deep convolutional networks for large-scale image recognition[J/OL]. (2015-04-10) [2021-09-17]. DOI: 10.1109/CVPR.2016.182.

11.Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[J/OL]. (2015-10-15) [2021-09-17]. DOI: 10.1109/CVPR.2015.7298594. 

12.He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[J/OL]. (2016-12-12) [2021-09-17]. DOI: 10.1109/CVPR.2016.90.

13.增思涛, 曹永春,林强, 等. 基于ResNet深度模型的 SPECT肺灌注图像分类[J]. 西 北 民 族 大 学 学 报(自 然 科 学 版), 2020, 42(2): 27-35. [Zeng ST, Cao YC, Lin Q, et al. Classifying SPECT lung perfusion images based on ResNet models[J]. Journal of Northwest University for Nationalities (Natural Science), 2020, 42(2): 27-35.] DOI: 10.3969/j.issn.1009-2102.2021.02.006.

14.韩晓臻, 金冉, 李家辉, 等. 一种基于ResNet与迁移学习的小样本图像识别方法[J]. 浙江万里学院学报, 2020, 33(6): 82-90. [Han XZ, Jin R, Li JH, et al. A small sample image recognition method based on ResNet and transfer learning[J]. Journal of Zhejiang Wanli University, 2020, 33(6): 82-90.] DOI: 10.13777/j.cnki.issn1671-2250.2020.06.014.

15.Becker AS, Mueller M, Stoffel E, et al. Classification of breast cancer in ultrasound imaging using a generic deep learning analysis software: a pilot study[J]. Br J Radiol, 2018, 91(1083): 20170576. DOI: 10.1259/bjr.20170576.

16.常冬霞, 王舒伟. 基于鲁棒回归度量学习的图像分类算法[J]. 北京交通大学学报, 2021, 45(2): 119-126.  [Chang DX, Wang SW. Robust regression metric learning algorithm for image classification[J]. Journal of Bei-jing Jiaotong University, 2021, 45(2): 119-126.] DOI: 10.11860/j.issn.1673-0291.20200134.

Hot Papers