肺炎是世界上造成五歲以下小孩主要死亡原因之一。目前，臨床上透過胸腔 X 光影像進行診斷，由於影像病灶不清楚導致容易被誤診，因此相當仰賴放射科醫師的臨床經驗。由於卷積神經網路與遷移學習方法在電腦輔助診斷上的成功，使得數據集少的醫學影像在檢測疾病中得以有突破。因此，本研究希望可以藉由微調卷積神經網路輔助經驗不足的醫生，或是協助偏遠地區的診所診斷肺炎或是肺炎種類。
在這項研究中，使用著名的圖像分類卷積神經網路 VGG16、VGG19 以及 Inception V3，模型皆使用大數據 ImageNet 預訓練完成。藉由影像中初級特徵的通用性，並接上設計的分類器，從胸腔 X 光影像中檢測肺炎與肺炎種類。最終，三種模型在檢測肺炎的準確率高達98%，並且在 VGG16 中區分細菌性肺炎與病毒性肺炎的準確率達到84%。實驗結果證明先判斷有無肺炎，再區分肺炎種類可以提升診斷的正確性，足夠成為臨床醫師在診斷肺炎時的依據。

Pneumonia is among the top diseases which cause most of the deaths in children all over the world. Nowadays, the disease can be diagnosed from chest X-ray images, but it may be easily misdiagnosed due to unclear imaging lesions. Therefore, it depends on the clinical experience of radiologists. Recently, the success of Convolutional Neural Networks (CNNs) in the application of medical imaging and the emergence of transfer learning, medical imaging with few data sets has made a breakthrough in the detection of diseases. Therefore, the purpose of this study is that computer-aided diagnosis systems can assist inexperienced doctors or clinics in remote areas to diagnose pneumonia or types of pneumonia.
In this study, the well-known Convolutional Neural Networks were used, including VGG16, VGG19 and Inception V3. The models were pre-trained through ImageNet. With the versatility of the primary features in the image and the designed classifier, the pneumonia and pneumonia types are detected from the chest X-ray image. In the end, The test results showed that the accuracy of the models in diagnosis pneumonia was 98%, and the accuracy of distinguishing bacterial pneumonia from viral pneumonia in VGG16 achieved 84%. The results prove that judging whether patient suffer pneumonia or not before distinguishing the types of pneumonia can improve the accuracy of pneumonia diagnosis, which is enough to be the basis for clinicians when diagnosing pneumonia.

參考文獻
[1] Mathur, S., Fuchs, A., Bielicki, J., Van Den Anker, J., & Sharland, M. (2018). Antibiotic use for community-acquired pneumonia in neonates and children: WHO evidence review. Paediatrics and international child health, 38(sup1), S66–S75.
[2] Aydoğdu, M., Ozyilmaz, E., Aksoy, H., Gürsel, G., & Ekim, N. (2010). Mortality prediction in community-acquired pneumonia requiring mechanical ventilation; values of pneumonia and intensive care unit severity scores. Tuberkuloz ve toraks, 58(1), 25–34.
[3] World Health Organization. Pneumonia Vaccine Trial Investigators'
Group & World Health Organization. (2001). Standardization of
interpretation of chest radiographs for the diagnosis of pneumonia in
children / World Health Organization Pneumonia Vaccine Trial
Investigators' Group.
[4] Ruuskanen, O; Lahti, E; Jennings, LC; Murdoch, DR. Viral
pneumonia. (2011). Lancet, 377 (9773): 1264–75.
[5] Hopstaken, R. M., Witbraad, T., van Engelshoven, J. M., & Dinant,
G. J. (2004). Inter-observer variation in the interpretation of chest
radiographs for pneumonia in community-acquired lower respiratory
tract infections. Clinical radiology, 59(8), 743–752.
[6] Mahdieh Poostchi, Kamolrat Silamut, Richard Maude, Stefan
Jaeger, and George Thoma. (2018). Image analysis and machine
learning for detecting malaria. Translational Research, 194, 36-55.
[7] Harry Pratta, Frans Coenenb, Deborah M.Broadbentc, Simon
P.Hardingac, Yalin Zhengac. (2016). Convolutional Neural Networks
for Diabetic Retinopathy. Procedia Computer Science, 90, 200-205.
[8] Y. Fujisawa, Y. Otomo, Y. Ogata, Y. Nakamura, R. Fujita, Y.
Ishitsuka. (2019). Deep-learning-based, computer-aided classifier
developed with a small dataset of clinical images surpasses boardcertified dermatologists in skin tumour diagnosis. Janda and Soyer. Br J Dermatol, 180, 247–248.
[9] Dinggang Shen, Guorong Wu, and Heung-Il Suk. (2017). Deep
Learning in Medical Image Analysis. Biomed, 19, 21–48.
[10]G. Litjens et al. (2017). A survey on deep learning in medical image analysis. Medical image analysis, 42, 60-88.
[11]Ravishankar, H., Sudhakar, P., Venkataramani, R., Thiruvenkadam,
S., Annangi, P., Babu, N., & Vaidya, V. (2016). Understanding the
Mechanisms of Deep Transfer Learning for Medical Images. LABELS/DLMIA@MICCAI.
[12]Antin, B., Kravitz, J., & Martayan, E. (2017). Detecting Pneumonia in Chest X-Rays with Supervised Learning.
[13]Paras Lakhani and Baskaran Sundaram. (2017). Deep learning at
chest radiography: automated classification of pulmonary
tuberculosis by using convolutional neural networks. Radiology,
284(2), 574-582.
[14]Wang, X., Peng, Y., Lu, L., Lu, Z., Bagheri, M., & Summers, R.M.
(2017). ChestX-Ray8: Hospital-Scale Chest X-Ray Database and
Benchmarks on Weakly-Supervised Classification and Localization
of Common Thorax Diseases. 2017 IEEE Conference on Computer
Vision and Pattern Recognition (CVPR), 3462-3471.
[15]Rajpurkar, P., Irvin, J., Zhu, K., Yang, B., Mehta, H., Duan, T., Ding, D.Y., Bagul, A., Langlotz, C.P., Shpanskaya, K.S., Lungren, M.P., & Ng, A.Y. (2017). CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning. ArXiv, abs/1711.05225.
[16]D. Varshni, K. Thakral, L. Agarwal, R. Nijhawan and A. Mittal.
(2019). Pneumonia Detection Using CNN based Feature
Extraction. 2019 IEEE International Conference on Electrical,
Computer and Communication Technologies (ICECCT), 1-7.
[17]Glorot, X., Bordes, A. & Bengio, Y. (2011). Deep Sparse Rectifier
Neural Networks. Proceedings of the Fourteenth International
Conference on Artificial Intelligence and Statistics, in PMLR
15:315-323
[18]Ramachandran, P., Zoph, B., & Le, Q.V. (2017). Swish: a Self-Gated Activation Function. arXiv: Neural and Evolutionary Computing.
[19]Zhang, Z., & Sabuncu, M.R. (2018). Generalized Cross Entropy
Loss for Training Deep Neural Networks with Nois Labels. NeurIPS.
[20]Bottou, L., Curtis, F.E., & Nocedal, J. (2018). Optimization Methods for Large-Scale Machine Learning. SIAM Review, 60, 223-311.
[21]Suvrit Sra, Sebastian Nowozin, and Stephen J. Wright. (2011).
Optimization for Machine Learning. The MIT Press.
[22]Kingma, D.P., & Ba, J. (2015). Adam: A Method for Stochastic
Optimization. CoRR, abs/1412.6980.
[23]Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., &
LeCun, Y. (2014). OverFeat: Integrated Recognition, Localization
and Detection using Convolutional Networks. CoRR, abs/1312.6229.
[24]Y. Lecun, L. Bottou, Y. Bengio and P. Haffner. (1998). Gradientbased learning applied to document recognition. in Proceedings of the IEEE, 86(11), 2278-2324
[25]Khan, A., Sohail, A., Zahoora, U., & Qureshi, A.S. (2020). A survey of the recent architectures of deep convolutional neural
networks. Artificial Intelligence Review, 1 - 62.
[26]Krizhevsky, Alex & Sutskever, Ilya & Hinton, Geoffrey. (2012).
ImageNet Classification with Deep Convolutional Neural Networks.
Neural Information Processing Systems. 25. 10.1145/3065386.
[27]Simonyan, K., & Zisserman, A. (2015). Very Deep Convolutional
Networks for Large-Scale Image Recognition. CoRR, abs/1409.1556.
[28]Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015). Going deeper
with convolutions. 2015 IEEE Conference on Computer Vision and
Pattern Recognition (CVPR), 1-9.
[29]Lin, M., Chen, Q., & Yan, S. (2014). Network In Network. CoRR,
abs/1312.4400.
[30]Ioffe, S., & Szegedy, C. (2015). Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. ArXiv, abs/1502.03167.
[31]Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z.
(2016). Rethinking the Inception Architecture for Computer
Vision. 2016 IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), 2818-2826.
[32]S. J. Pan and Q. Yang. (2010). A Survey on Transfer Learning. in
IEEE Transactions on Knowledge and Data Engineering, 22,(10),
1345-1359.
[33]Torrey, Lisa, and Jude Shavlik. (2010). Transfer learning. In
Handbook of research on machine learning applications and trends:
algorithms, methods, and techniques, 242-264.
[34]Yosinski, J., Clune, J., Bengio, Y., & Lipson, H. (2014). How
transferable are features in deep neural networks? ArXiv,
abs/1411.1792.
[35]Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H.,
Laviolette, F., Marchand, M., & Lempitsky, V.S. (2016). DomainAdversarial Training of Neural Networks. ArXiv, abs/1505.07818.
[36]Johnson, M., Schuster, M., Le, Q.V., Krikun, M., Wu, Y., Chen, Z., Thorat, N., Viégas, F.B., Wattenberg, M., Corrado, G.S., Hughes,
M., & Dean, J. (2017). Google’s Multilingual Neural Machine
Translation System: Enabling Zero-Shot Translation. Transactions of
the Association for Computational Linguistics, 5, 339-351.
[37]Kermany, D. S., Goldbaum, M., Cai, W., Valentim, C., Liang, H.,
Baxter, S. L., McKeown, A., Yang, G., Wu, X., Yan, F., Dong, J.,
Prasadha, M. K., Pei, J., Ting, M., Zhu, J., Li, C., Hewett, S., Dong, J., Ziyar, I., Shi, A., … Zhang, K. (2018). Identifying Medical
Diagnoses and Treatable Diseases by Image-Based Deep
Learning. Cell, 172(5), 1122–1131.e9.
[38]Shorten, C., & Khoshgoftaar, T.M. (2019). A survey on Image Data
Augmentation for Deep Learning. Journal of Big Data, 6, 1-48.
[39]Tom Fawcett. (2006). An introduction to ROC analysis. Pattern
Recogn. Lett. 27, 8, 861–874.
[40]Bradley, Andrew P. (1997) The use of the area under the ROC curve
in the evaluation of machine learning algorithms. Pattern
Recognition, 30(7), 1145-1159.