肺癌是導致癌症死亡的常見原因之一，相較於傳統的胸部X光檢查，使用電腦斷層掃描(Computed Tomography，CT)進行的肺癌篩檢可將肺癌死亡率降低20％。因此，肺部CT圖像檢測在全世界廣泛使用。近年來，深度學習中的卷積神經網路圖像處理上展現出比傳統辨識方法更加優異的效能。卷積神經網路將影像以數百萬的神經網路參數向輸出端傳遞，於輸出端計算目標與預測的誤差，達成具高變化且高維的影像辨識。因而有越來越多研究嘗試以不同的深度學習技術，希望提高肺癌影像檢測準確率。
因此本研究提出以田口法參數設計優化卷積神經網路(Taguchi-based Convolutional neural network,T-CNN)，對CT圖像進行肺癌辨識分類。在田口法中，以8個控制因子3個水準選擇了L36進行實驗來確定T-CNN的最佳參數。本研究以兩個肺癌數據集分別為(A) TCIA（LIDC-IDRI）和(B) TCIA (SPIE-AAPM Lung CT Challenge)進行實驗。實驗結果證明採用田口參數優化的T-CNN和原始CNN在肺癌辨識的平均分類準確率，數據集(A) 分別為98.83％和91.97％；數據集(B)分別為99.97％和94.68％，比原始的CNN分別提高了6.86％和5.29％，證實本研究提出一個更高的準確度模型，並提高了肺癌的分類準確性。

Lung cancer is one of the common causes of cancer death. Compared with traditional chest X-ray examination, lung cancer screening using Computed Tomography (CT) can reduce lung cancer mortality by 20%. Therefore, lung CT image detection is widely used all over the world. In recent years, the Convolutional Neural Network (CNN) in deep learning (DNN) has shown more excellent performance in image processing than traditional identification methods. The convolutional neural network transfers the image to the output with millions of neural network parameters, and calculates the error between the target and the prediction at the output to achieve high-variation and high-dimensional image recognition.
Therefore, this study proposes to use the Taguchi method to design and optimize the Taguchi-based Convolutional neural network (T-CNN) to classify the CT images for lung cancer identification. In the Taguchi method, L36 was selected with 8 control factors and 3 levels to conduct experiments to determine the optimal parameters of T-CNN. In this study, the two lung cancer data sets were (A) TCIA (LIDC-IDRI) and (B) TCIA SPIE-AAPM Lung CT Challenge. The experimental results prove that the average classification accuracy of T-CNN and original CNN optimized by Taguchi parameters in lung cancer identification is 98.83% and 91.97% in data set (A); 99.97% and 94.67% respectively in data set (B) The original CNN was improved by 6.86% and 5.29% respectively, confirming that this study proposed a higher accuracy model and improved the classification accuracy of lung cancer.

摘要 I
ABSTRACT III
誌謝 V
目錄 VI
表目錄 VIII
圖目錄 X
第一章 緒論 1
1.1研究動機與背景 1
1.2研究目的 4
1.3 論文架構 5
第二章 文獻探討 7
2.1 卷積神經網路 7
2.2 肺部疾病相關研究 11
2.3 田口法 (Taguchi method)相關研究 12
第三章 研究方法 14
3.1田口法 15
3.2資料集來源 20
3.3卷積神經網路架構 27
3.4進行模型評估驗證 34
第四章 實驗結果 35
4.1 LIDC-IDRI資料集實驗結果 36
4.2 SPIE-AAPM資料集實驗結果 44
4.3 實驗結果分析 49
第五章 結論 52
5.1 結論 52
5.2 未來發展 54
參考文獻 55

[1]F.Bray, J.Ferlay, I.Soerjomataram, R.L.Siegel, L.A. Torre, A. Jemal, "Global Cancer Statistics 2018: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries, "CA: A Cancer Journal for Clinicians., Vol. 68, No. 6, pp. 394-424, 2018.
[2]G. Polat, Y. S. Dogrusoz, U. Halici, " Effect of Input Size on the Classification of Lung Nodules Using Convolutional Neural Networks. SIU (p./pp. 1-4), IEEE. ISBN: 978-1-5386-1501-0, 2018.
[3]P. V. Sankar and J. Sklansky, "A Gestalt-Guided Heuristic Boundary Follower for X-Ray Images of Lung Nodules," In IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-4, No. 3, pp. 326-331, 1982.
[4]Y. Hagihara ,Jun Wei and H. Kobatake, "Detection of Rounded Opacities on Chest Radiographs Using Convergence Index Filter," Proceedings 10th International Conference on Image Analysis and Processing, Venice, Italy, pp. 757-761, 1999.
[5]Y. S. P. Chiou, Y. M. F. Lure and P. A. Ligomenides, "Neural Network Image Analysis and Classification in Hybrid Lung Nodule Detection (HLND) System," Neural Networks for Signal Processing III - Proceedings of the 1993 IEEE-SP Workshop, Linthicum Heights, MD, pp. 517-526, 1993.
[6]J.S. Lin, P. A. Ligomenides, S.C. B. Lo, A. Hasegawa, M. T. Freedman and S. K. Mun, "An Application of Convolution Neural Networks: Reducing False-positives in Lung Nodule Detection," Proceedings of 1994 IEEE Nuclear Science Symposium - NSS'94, Norfolk, VA, Vol.4, pp. 1842-1846 ,1994.
[7]A. A. Abdullah and H. Mohamaddiah, "Development of Cellular Neural Network Algorithm for Detecting Lung Cancer Symptoms," 2010 IEEE EMBS Conference on Biomedical Engineering and Sciences (IECBES), Kuala Lumpur, pp. 138-143, 2010.
[8]P. Hattikatti, "Texture Based Interstitial Lung Disease Detection Using Convolutional Neural Network," 2017 International Conference on Big Data, IoT and Data Science (BID), Pune, pp. 18-22, 2017.
[9]Q. Li, W. Cai, X. Wang, Y. Zhou, D. D. Feng and M. Chen, "Medical Image Classification with Convolutional Neural Network," 2014 13th International Conference on Control Automation Robotics & Vision (ICARCV),Singapore, pp. 844-848, 2014.
[10]H. Liu, S. Zhang, X. Jiang, T. Zhang, H. Huang, F. G. Lin ,Z. Xiao, X. Li, J.H. Hu, G.Lei, T. Liu, "The Cerebral Cortex is Bisectionally Segregated into Two Fundamentally Different Functional Units of Gyri and Sulci, "Cerebral Cortex, Vol. 29, No. 10, pp.4238–4252, 2019.
[11]S.Zhang, H.Liu, H.Huang, Y.Zhao, X.Jiang, B.Bowers, L. Guo, X. Hu, M. Sanchez, T.Liu , "Deep Learning Models Unveiled Functional Difference between Cortical Gyri and Sulci," in IEEE Transactions on Biomedical Engineering, Vol. 66, No. 5, pp. 1297-1308, 2019.
[12]D. Riquelme,M. A. Akhloufi, "Deep Learning for Lung Cancer Nodules "Detection and Classification in CT Scans. AI 2020, Vol. 1, pp. 28-67,2020.
[13]Y. Zhao, G.H. D. Bock, R. Vliegenthart , R. J.V. Klaveren, Y.Wang, L. Bogoni, P. A. D.Jong, W.P.Mali, P.M.A.V.Ooijen, M.Oudkerk,"Performance of Computer-aided Detection of Pulmonary Nodules in Low-dose CT: Comparison with Double Reading by Nodule. "Eur .Radiol, Vol. 22, pp. 2076–2084, 2012.
[14]P. M. Arabi, N. V. Deepa, N. Krishna and T. P. Prathibha, "Identification of Ground Glass Structure in Lung CT Images by Computer Aided Diagnosis Technique," 2016 10th International Conference on Intelligent Systems and Control (ISCO), Coimbatore, pp. 1-4, 2016.
[15]C. Qin, D. Yao, Y. Shi, Z. Song, " Computer-aided Detection in Chest Radiography Based on Artificial Intelligence: a Survey. Biomed Eng., Online , Vol.17, No.1, pp.113, Doi:10.1186/s12938-018-0544-y, 2018.
[16]S. Suresh, S. Mohan, " NROI Based Feature Learning for Automated Tumor Stage Classification of Pulmonary Lung Nodules Using Deep Convolution.al Neural Networks. Journal of King Saud University –
" Computer and Information Sciences. DOI: https://doi.org/10.1016 /j.jksuci.2019.11.013, 2019.
[17]M. Yunus, M. S. Alsoufi, "A Statistical Analysis of Joint Strength of Dissimilar Aluminium Alloys Formed by Friction Stir Welding Using Taguchi Design Approach, ANOVA for the Optimization of Process Parameters. "International Journal of Research in Engineering & Technology. Vol. 3, No. 6, pp. 63-70, 2015.
[18]T. Canel, M. Zeren,T. Sınmazçelik, " Laser Parameters Optimization of Surface Treating of Al 6082-T6 with Taguchi Method. " Optics and Laser Technology. Vol. 120, https://doi.org/10.1016/j.optlastec.2019.105714 , 2019.
[19]A. Yang, X. Yang, W. Wu, H. Liu and Y. Zhuansun, "Research on Feature Extraction of Tumor Image Based on Convolutional Neural Network," IEEE Access, Vol. 7, pp. 24204-24213, 2019.
[20]M. H. Othman and S. Y. M. Amin, "Optimising Injection Moulding Parameters that Satisfies Part Qualities by Using Taguchi Method," 2012 IEEE Business, Engineering & Industrial Applications Colloquium (BEIAC), Kuala Lumpur, pp. 307-312, 2012.
[21]D. H. Hubel, T. N. Wiesel, "Receptive Fields, Binocular Interaction and Functional Aarchitecture in the Cat's Visual Cortex, "The Journal of Physiology , Vol. 160, Doi: 10.1113/jphysiol.1962.sp006837 , 1962.
[22]K. Fukushima and S. Miyake, "Neocognitron: A New Algorithm for Pattern Recognition Tolerant of Deformations and Shifts in Position, " Pattern Recognit., Vol.15, No.6, pp. 455 –469, 1982.
[23]D.Rumelhart,G.Hinton, R. Williams, "Learning Representations by Back-propagating Errors, " Nature ,Vol. 323, pp. 533–536 , 1986.
[24]W.Wang, Y.Yang, X.Wang, W.Wang, J. Li, " Development of Convolutional Neural Network and Its Application in Image Classification: a Survey, "Optical Engineering, Vol.58, https://doi.org /10.1117/1.OE.58.4.040901 , 2019.
[25]Y. Lecun, L. Bottou, Y. Bengio and P. Haffner, "Gradient-based Learning Applied to Document Recognition," In Proceedings of the IEEE, Vol. 86, No. 11, pp. 2278-2324, 1998.
[26]M. Jones and T. Poggio, "Regularization Theory and Neural Networks Architectures, " Neural Comp.,Vol.7, No.2, pp. 219–269, 1995.
[27]Alex Krizhevsky, I. Sutskever, E. G. Hinton , "International Conference on Neural Information Processing Systems – Vol.1,pp. 1097–1105, 2012.
[28] Stanford Vision Lab, Stanford University, Princeton University support@image-net.org . ©2016
[29]N.Srivastava, G.Hinton, A.Krizhevsky, I.Sutskever, R.Salakhutdinov, "Dropout: A Simple Way to Prevent Neural Networks from Overfitting, " J. Mach. Learn. Res., Vol. 15, pp. 1929–1958, 2014.
[30]K. Simonyan, A. Zisserman, "Very Deep Convolutional Networks for Large-Scale Image Recognition,"arXiv: 1409.1556, 2014.
[31]C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D.Erhan, V.Vanhoucke, A. Rabinovich, "Going Deeper with Convolutions, " arXiv: 1409.4842 , 2014.
[32]K.He, X. Zhang, S. Ren, J. Sun, "Deep Residual Learning for Image Recognition, " arXiv:1512.03385. 2015.
[33]X. Yu, Y. Fu and T. Liu, "Face Detection: A Deep Convolutional Network Method Based on Grouped Facial Part," 2017 IEEE 2nd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC), Chongqing, pp. 515-519, 2017.
[34]S. K. Pandey, H. S. Shekhawat and S. R. M. Prasanna, "Deep Learning Techniques for Speech Emotion Recognition: A Review," 2019 29th International Conference Radioelektronika (RADIOELEKTRONIKA), Pardubice Czech, pp. 1-6, 2019.
[35]R. A. Hamad, A. S. Hidalgo, M. Bouguelia, M. E. Estevez and J. M. Quero, "Efficient Activity Recognition in Smart Homes Using Delayed Fuzzy Temporal Windows on Binary Sensors," IEEE Journal of Biomedical and Health Informatics, Vol. 24, No. 2, pp. 387-395, 2020.
[36]Y. Qin, M. Chen, L. Qi and Y. Sun, "Focus Generator with Score Classification on Fabric Defect Detection," 2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI), Portland, OR, pp. 1708-1714, 2019.
[37]P. Casas, G. Marín, G. Capdehourat and M. Korczynski, "MLSEC - Benchmarking Shallow and Deep Machine Learning Models for Network Security," 2019 IEEE Security and Privacy Workshops (SPW), San Francisco, CA, pp. 230-235, 2019.
[38]T. Do, M. Duong, Q. Dang and M. Le, "Real-Time Self-Driving Car Navigation Using Deep Neural Network," 2018 4th International Conference on Green Technology and Sustainable Development (GTSD), Ho Chi Minh City, pp. 7-12, 2018.
[39]R. Yamashita, M. Nishio, R.K.G. Do, K. Togashi, " Insights Imaging " Convolutional neural networks: an overview andapplication in radiology. Vol.9, pp.611–629,2018.
[40]A.Masood, B.Sheng, P.Li, X.Hou, X.Wei, J.Qin, D.Feng, "Computer-assisted Decision Support System in Pulmonary Cancer Detection and Stage Classification on CT Images. "Journal of Biomedical Informatics. Vol.79, pp. 117-128, 2018.
[41]A.O.D. C.Filho,W.B.D. Sampaio, A.C.Silva,A.C.Silva, A.C.D. Paiva, R. A.Nunes, M.Gattass, "Automatic Detection of Solitary Lung Nodules Using Quality Threshold Clustering, Genetic Algorithm and Diversity Index , "Artificial Intelligence in Medicine, Vol. 60, No. 3, pp. 165-177, 2014.
[42]廖文丕,「運用田口法找出以靈芝大豆發酵液多醣體目標值最佳化製程之研究」碩士論文，國立勤益科技大學研發科技與資訊管理研究所, 2010.
[43]P. Sonawane, V. B. Savakhande, M. A. Chewale and R. A. Wanjari, "Optimization of PID Controller for Automatic Voltage Regulator System Using Taguchi Method," 2018 International Conference on Computer Communication and Informatics (ICCCI), Coimbatore, pp. 1-6, 2018.
[44]C. Lee, B. Su, C. Chang, T. Hsu and W. Lee, "Applications of Taguchi Method to PID Control for Path Tracking of A Wheeled Mobile Robot," 2018 IEEE International Conference on Applied System Invention (ICASI), Chiba, pp. 453-456, 2018.
[45]A.Yaser,A. J.J. Cabibihan, A.Hayek, A.Bonarini, " Safety Experiments for Small Robots Investigating the Potential of Soft Materials in Mitigating the Harm to the Head Due to Impacts , "arXiv:1904.09621, 2019.
[46]A. Bhole,S. H.Turlapati,V. S. Rajashekhar,J. Dixit,S.V. Shah,K M. Krishna , "Design of a Robust Stair Climbing Compliant Modular Robot to Tackle Overhang on Stairs, "arXiv:1607.03077, 2016.
[47]戴搖廷,應用田口方法至 TFT-LCD 黑色矩陣檢測缺陷之參數設計,國立中興大學電機工程學系所碩士論文，台中市, https://hdl.handle.net/11296/7fn76m , 2014.
[48]盧昆宏,江季哲,模糊田口法於多重品質特性製程上之研究--以TFT金屬鍍膜製程為例, 品質學報 17:1.02[民99.02] 頁67-87,2010.
[49]F.N. Idris, M.M. Nadzir, S.R.A. Shukor, " Optimization of Solvent-free Microwave Extraction of Centella Asiatica Using Taguchi Method. "Journal of Environmental Chemical Engineering, Available Online, DOI: https://doi.org /10.1016 /j.jece.2020.103766, 2020.
[50]張榮宗,利用田口方法探討腹部X光攝影條件之最佳化,中臺科技大學醫學影像暨放射科學系暨研究所碩士論文，台中市, https://hdl.handle.net/11296/fx988u, 2014.
[51]永析,單因子獨立變異數分析 (Analysis of Variance, ANOVA)-統計說明與SPSS操作https:// www. Yongxi -stat.com /one-way-anova -indenpedent /,6月2017.
[52]李阮英武,添加鋁奈米粒子潤滑油之潤滑性質研究,國立高雄應用科技大學機械工程系博士論文,高雄市, https://hdl.handle.net /11296/t34s7r, 2017.
[53]The Cancer Imaging Archive,TCIA.https://www.Cancerimagingar chive.net /about-the-cancer- imaging -archive-tcia/.
[54]G.Samuel,Armato III, M. Geoffrey, B. Luc, M.Gray, F.Michael, M. R. Charles, R. P. Anthony, L. P. Clarke,Data From LIDC-IDRI. " The Cancer Imaging Archive , http://doi.org/10.7937/K9/TCIA.2015.LO9QL9SX, 2015.
[55]G.Samuel ,Armato III; H. T. G. D. Drukker, M. Li, F. Redmond, S.Justin, P.L.Clarke, From SPIE-AAPM-NCI Lung Nodule Classification Challenge Dataset. "The Cancer Imaging Archive.http://doi.org/10.7937/K9/ TCIA.2015.UZLSU3FL, 2015.
[56]T.Xiao,L.Liu,K.Li,W.Qin,S.Yu,Z.Li, " Comparison of Transferred Deep Neural Networks in Ultrasonic Breast Masses Discrimination https:// doi.org/10.1155/2018/ 4605191 , 2018.
[57]C.J.Lin,S.Y.Jeng,M.K.Chen ,"Using 2D CNN with Taguchi Parametric Optimization for Lung Cancer Recognition from CT Images, "Appl. Sci. 10(7), 2591;https://doi.org/10.3390 /app10072591, 2020.
[58]M. Togaçar, B. Ergen, Z. Cömert, " Detection of Lung Cancer on Chest CT Images Using Minimum Redundancy Maximum Relevance Feature Selection Method with Convolutional Neural Networks. "Biocybernetics and Biomedical Engineering. Vol. 40, No. 1, pp. 23-39, 2020.
[59]E.E. Nithila, S.S. Kumar, "Automatic Detection of Solitary Pulmonary Nodules Using Swarm Intelligence Optimized Neural Networks on CT Images. "Engineering Science and Technology, an International Journal. Vol. 20, pp. 1192-1202, 2017.