Journal of the Korean Society of Radiology (대한영상의학회지)

The Korean Society of Radiology (대한영상의학회)
권호 표지
DOI QR코드

DOI QR Code

Development of an Optimized Deep Learning Model for Medical Imaging

의료 영상에 최적화된 딥러닝 모델의 개발

  • Young Jae Kim (Department of Biomedical Engineering, Gachon University) ;
  • Kwang Gi Kim (Department of Biomedical Engineering, Gachon University)
  • 김영재 (가천대학교 의용생체공학과) ;
  • 김광기 (가천대학교 의용생체공학과)
  • Received : 2020.10.03
  • Accepted : 2020.11.02
  • Published : 2020.11.01
Download PDF
⟨ Previous Next ⟩

Abstract

Deep learning has recently become one of the most actively researched technologies in the field of medical imaging. The availability of sufficient data and the latest advances in algorithms are important factors that influence the development of deep learning models. However, several other factors should be considered in developing an optimal generalized deep learning model. All the steps, including data collection, labeling, and pre-processing and model training, validation, and complexity can affect the performance of deep learning models. Therefore, appropriate optimization methods should be considered for each step during the development of a deep learning model. In this review, we discuss the important factors to be considered for the optimal development of deep learning models.

최근, 의료 영상 분야에서 딥러닝은 가장 활발하게 연구되고 있는 기술 중 하나이다. 충분한 데이터와 최신의 딥러닝 알고리즘은 딥러닝 모델의 개발에 중요한 요소이다. 하지만 일반화된 최적의 딥러닝 모델을 개발하기 위해서는 데이터의 양과 최신의 딥러닝 알고리즘 외에도 많은 것을 고려해야 한다. 데이터 수집부터 가공, 전처리, 모델의 학습 및 검증, 경량화까지 모든 과정이 딥러닝 모델의 성능에 영향을 미칠 수 있기 때문이다. 본 종설에서는 의료 영상에 최적화된 딥러닝 모델을 위해 개발 과정 각각에서 고려해야 할 중요한 요소들을 살펴보고자 한다.

Keywords

Acknowledgement

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2019R1C1C1008381), and Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government (MSIT) (No. 2019-0-01750-002, Development of an optimal limb-compressing cardiovascular treatment device using deep learning technique) (No. 2020-0-00161-001, Active Machine Learning based on Open-set training for Surgical Video).

References

  1. Amato F, Lopez A, Pena-Mendez EM, Vanhara P, Hampl A, Havel J. Artificial neural networks in medical diagnosis. J Appl Biomed 2013;11:47-58 https://doi.org/10.2478/v10136-012-0031-x
  2. Ren J. ANN vs. SVM: which one performs better in classification of MCCs in mammogram imaging. Knowl Based Syst 2012;26:144-153 https://doi.org/10.1016/j.knosys.2011.07.016
  3. Cinar M, Engin M, Engin EZ, Atesci YZ. Early prostate cancer diagnosis by using artificial neural networks and support vector machines. Expert Syst Appl 2009;36:6357-6361 https://doi.org/10.1016/j.eswa.2008.08.010
  4. Hinton GE, Salakhutdinov RR. Reducing the dimensionality of data with neural networks. Science 2006;313:504-507 https://doi.org/10.1126/science.1127647
  5. Hinton GE, Srivastava N, Krizhevsky A, Sutskever I, Salakhutdinov RR. Improving neural networks by preventing co-adaptation of feature detectors. ArXiv Preprint 2012;arXiv:1207.0580
  6. Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. New York: Communications of the ACM 2012:1-9
  7. Silver D, Huang A, Maddison CJ, Guez A, Sifre L, Van den Driessche G, et al. Mastering the game of go with deep neural networks and tree search. Nature 2016;529:484-489 https://doi.org/10.1038/nature16961
  8. Tajbakhsh N, Shin JY, Gurudu SR, Hurst RT, Kendall CB, Gotway MB, et al. Convolutional neural networks for medical image analysis: full training or fine tuning? IEEE Trans Med Imaging 2016;35:1299-1312 https://doi.org/10.1109/TMI.2016.2535302
  9. Chen J, Wang Y, Wu Y, Cai C. An ensemble of convolutional neural networks for image classification based on LSTM. Proceedings of 2017 International Conference on Green Informatics (ICGI); 2017 Aug 15-17; Fuzhou, China: ICGI; 2017:217-222
  10. Anwar SM, Majid M, Qayyum A, Awais M, Alnowami M, Khan MK. Medical image analysis using convolutional neural networks: a review. J Med Syst 2018;42:226
  11. Wang Z, Yan W, Oates T. Time series classification from scratch with deep neural networks: a strong baseline. Proceedings of IJCNN 2017 : International Joint Conference on Neural Networks; 2017 May 14-19; Anchorage, AK, USA: IJCNN; 2017:1578-1585
  12. Philbrick KA, Yoshida K, Inoue D, Akkus Z, Kline TL, Weston AD, et al. What does deep learning see? Insights from a classifier trained to predict contrast enhancement phase from CT images. AJR Am J Roentgenol 2018;211:1184-1193 https://doi.org/10.2214/AJR.18.20331
  13. Bona JP, Prior FW, Zozus MN, Brochhausen M. Enhancing clinical data and clinical research data with biomedical ontologies - Insights from the knowledge representation perspective. Yearb Med Inform 2019;28:140-151 https://doi.org/10.1055/s-0039-1677912
  14. Basu A, Warzel D, Eftekhari A, Kirby JS, Freymann J, Knable J, et al. Call for data standardization: lessons learned and recommendations in an imaging study. JCO Clin Cancer Inform 2019;3:1-11 https://doi.org/10.1200/CCI.19.00056
  15. Aberle DR, Berg CD, Black WC, Church TR, Fagerstrom RM, Galen B, et al. The National Lung Screening Trial: overview and study design. Radiology 2011;258:243-253 https://doi.org/10.1148/radiol.10091808
  16. Yan K, Wang X, Lu L, Zhang L, Harrison AP, Bagheri M, et al. Deep lesion graphs in the wild: relationship learning and organization of significant radiology image findings in a diverse large-scale lesion database. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; 2018 Jun 18-22; Salt Lake City, UT, USA: IEEE; 2018:9261-9270
  17. Wang X, Peng Y, Lu L, Lu Z, Bagheri M, Summers RM. Chestx-ray8: hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR); 2017 Jul 21-26; Honolulu, HI, USA: IEEE; 2017:2097-2106
  18. LaMontagne PJ, Benzinger TL, Morris JC, Keefe S, Hornbeck R, Xiong C, et al. OASIS-3: longitudinal neuroimaging, clinical, and cognitive dataset for normal aging and Alzheimer disease. MedRxiv 2019 [in press] doi: https://doi.org/10.1101/2019.12.13.19014902
  19. Benndorf M, Herda C, Langer M, Kotter E. Provision of the DDSM mammography metadata in an accessible format. Med Phys 2014;41:051902
  20. Yu S, Xiao D, Frost S, Kanagasingam Y. Robust optic disc and cup segmentation with deep learning for glaucoma detection. Comput Med Imaging Graph 2019;74:61-71 https://doi.org/10.1016/j.compmedimag.2019.02.005
  21. Schneider CA, Rasband WS, Eliceiri KW. NIH image to ImageJ: 25 years of image analysis. Nat Methods 2012;9:671-675 https://doi.org/10.1038/nmeth.2089
  22. Yushkevich PA, Gao Y, Gerig G. ITK-SNAP: an interactive tool for semi-automatic segmentation of multi-modality biomedical images. Proceedings of the 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); 2016 Aug 17-20; Orlando, FL, USA: IEEE; 2016:3342-3345
  23. Russell BC, Torralba A, Murphy KP, Freeman WT. LabelMe: a database and web-based tool for image annotation. Int J Comput Vis 2008;77:157-173 https://doi.org/10.1007/s11263-007-0090-8
  24. Hansen US, Landau E, Patel M, Hayee BH. Novel artificial intelligence-driven software significantly shortens the time required for annotation in computer vision projects. MedRxiv 2020 [in press] doi: https://doi.org/10.1101/2020.09.11.20192500
  25. Liu H, Bielinski SJ, Sohn S, Murphy S, Wagholikar KB, Jonnalagadda SR, et al. An information extraction framework for cohort identification using electronic health records. AMIA Jt Summits Transl Sci Proc 2013;2013:149-153
  26. Mehta R, Arbel T. 3D U-Net for brain tumour segmentation. In International MICCAI Brainlesion Workshop. Cham: Springer 2018:254-266
  27. Isensee F, Kickingereder P, Wick W, Bendszus M, Maier-Hein KH. No new-net. In International MICCAI Brainlesion Workshop. Cham: Springer 2018:234-244
  28. Rahman T, Khandakar A, Kadir MA, Islam KR, Islam KF, Mazhar R, et al. Reliable tuberculosis detection using chest X-ray with deep learning, segmentation and visualization. IEEE Access 2020;8:191586-191601 https://doi.org/10.1109/ACCESS.2020.3031384
  29. Yang S, Kweon J, Roh JH, Lee JH, Kang H, Park LJ, et al. Deep learning segmentation of major vessels in X-ray coronary angiography. Sci Rep 2019;9:13897
  30. Antoniou A, Storkey A, Edwards H. Data augmentation generative adversarial networks. ArXiv Preprint 2017;arXiv:1711.04340
  31. Han C, Murao K, Noguchi T, Kawata Y, Uchiyama F, Rundo L, et al. Learning more with less: conditional PGGAN-based data augmentation for brain metastases detection using highly-rough annotation on MR images. Proceedings of the 28th ACM International Conference on Information and Knowledge Management; 2019 Nov 3-7; Beijing, China: CIKM; 2019:119-127
  32. Sandfort V, Yan K, Pickhardt PJ, Summers RM. Data augmentation using generative adversarial networks (CycleGAN) to improve generalizability in CT segmentation tasks. Sci Rep 2019;9:16884
  33. Frid-Adar M, Diamant I, Klang E, Amitai M, Goldberger J, Greenspan H. GAN-based synthetic medical image augmentation for increased CNN performance in liver lesion classification. Neurocomputing 2018;321:321-331 https://doi.org/10.1016/j.neucom.2018.09.013
  34. Xin B, Hu Y, Zheng Y, Liao H. Multi-modality generative adversarial networks with tumor consistency loss for brain MR image synthesis. Proceedings of 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI); 2020 Apr 3-7; Iowa City, IA, USA: IEEE; 2020:1803-1807
  35. Ning CY, Liu SF, Qu M. Research on removing noise in medical image based on median filter method. Proceedings of 2009 IEEE International Symposium on IT in Medicine & Education (ITME2009); 2009 Aug 14-16; Jinan, China: IEEE; 2009:384-388
  36. Cadena L, Zotin A, Cadena F, Korneeva A, Legalov A. Noise reduction techniques for processing of medical images. Proceedings of the World Congress on Engineering 2017; Jul 5-7; London, UK: World Congress on Engineering; 2017:5-9
  37. Wang Q, Chen L, Shen D. Fast histogram equalization for medical image enhancement. Proceedings of 2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society; 2008 Aug 20-25; Vancouver, BC, Canada: IEEE; 2008:2217-2220
  38. Reza AM. Realization of the contrast limited adaptive histogram equalization (CLAHE) for real-time image enhancement. J Signal Process Sys 2004;38:35-44 https://doi.org/10.1023/B:VLSI.0000028532.53893.82
  39. Hashemi M. Enlarging smaller images before inputting into convolutional neural network: zero-padding vs. interpolation. J Big Data 2019;6:98
  40. Smith LN. A disciplined approach to neural network hyper-parameters: part 1 -- learning rate, batch size, momentum, and weight decay. ArXiv Preprint 2018;arXiv:1803.09820
  41. Shin HC, Roth HR, Gao M, Lu L, Xu Z, Nogues I, et al. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE Trans Med Imaging 2016;35:1285-1298 https://doi.org/10.1109/TMI.2016.2528162
  42. Shi Z, Hao H, Zhao M, Feng Y, He L, Wang Y, et al. A deep CNN based transfer learning method for false positive reduction. Multimed Tools Appl 2019;78:1017-1033 https://doi.org/10.1007/s11042-018-6082-6
  43. Wong TT, Yeh PY. Reliable accuracy estimates from k-fold cross validation. IEEE Trans Knowl Data Eng 2020;32:1586-1594 https://doi.org/10.1109/TKDE.2019.2912815
  44. Bleeker SE, Moll HA, Steyerberg EW, Donders AR, Derksen-Lubsen G, Grobbee DE, et al. External validation is necessary in prediction research: a clinical example. J Clin Epidemiol 2003;56:826-832 https://doi.org/10.1016/S0895-4356(03)00207-5
  45. Szegedy C, Ioffe S, Vanhoucke V, Alemi A. Inception-v4, inception-resnet and the impact of residual connections on learning. ArXiv Preprint 2016;arXiv:1602.07261
  46. Iandola F, Moskewicz M, Karayev S, Girshick R, Darrell T, Keutzer K. Densenet: implementing efficient convnet descriptor pyramids. ArXiv Preprint 2014;arXiv:1404.1869
  47. Iandola FN, Han S, Moskewicz MW, Ashraf K, Dally WJ, Keutzer K. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size. ArXiv Preprint 2016;arXiv:1602.07360
  48. Li Y, Huang H, Xie Q, Yao L, Chen Q. Research on a surface defect detection algorithm based on MobileNet-SSD. Appl Sci 2018;8:1678
  49. Zhang X, Lin M, Sun J. Shufflenet: an extremely efficient convolutional neural network for mobile devices. Proceedings of the IEEE conference on computer vision and pattern recognition; 2018 Jun 19-21; Salt Lake City, UT, USA: IEEE; 2018:6848-6856
  50. Han S, Mao H, Dally WJ. Deep compression: compressing deep neural networks with pruning, trained quantization and huffman coding. ArXiv Preprint 2015;arXiv:1510.00149
  51. Chen G, Choi W, Yu X, Han T, Chandraker M. Learning efficient object detection models with knowledge distillation. Proceedings of Advances in Neural Information Processing Systems; 2017 Dec 4-7; Long Beach, CA, USA: NIPS; 2017:1-10
  52. Yim J, Joo D, Bae J, Kim J. A gift from knowledge distillation: fast optimization, network minimization and transfer learning. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition; 2017 Jul 21-26; Honolulu, HI, USA: IEEE; 2017:4133-4141