臺灣博碩士論文加值系統

糖尿病視網膜病變是發展中國家勞動人口失明的主要原因。對此疾病的診斷嚴重依賴於影像判讀，且藉由訓練有素的臨床醫生進行的判讀既耗時且繁雜。如果能提高檢測過程的準確性和速度並用於早期診斷和干預，則能對人類社會健康產生重大影響。除了預防外，如何追蹤糖尿病視網膜病變患者的治療效果是個體化醫療保健中的另一個關鍵問題。受此啟發，我們提出了一種基於深度卷積神經網絡的識別框架。所提出的識別系統不僅可以預測糖網的嚴重程度，還可以預測像素級別的症狀位置。通過結合糖網嚴重程度水平和擷取的病症，系統可以更準確地預測糖網病變的嚴重程度水平，並提供另一種潛在的參考測量數值，以反應治療後糖網病變嚴重程度的演變。對於糖網病變嚴重性級別的分類辨識，所提出的輕量級網絡架構DRNet-cla-v1從兩個方面改進了分類性能：（1）在沒有重新訓練下，DRNet-cla-v1與其他七種增強方法相結合，在Messidor數據集中，分別針對DR 及RDR達到0.961和0.967的AUROC，其性能優於現有技術（0.921和0.957）。（2）與CKML Net，VNXK和Zoom-in-Net相比，DRNet-cla-v1的記憶體使用效率更高，總參數減少至少5.23倍，並只需要更低的計算成本，總FLOPs至少減少1.24倍。對於糖網病症的擷取，所提出網絡架構DRNet-seg-v1在IDRiD測試數據集上達到0.6894平均AUPRC，並且優於現有技術（0.6693）。最後，我們使用線性支持向量機融合從DRNet-cla-v1和DRNet-seg-v1中所提取的特徵，並在IDRiD測試數據集上實現平均準確度0.7281，也優於現有技術（0.6311）。

Diabetic retinopathy is the primary cause of blindness in the working-age population of the developed world. Diagnosing the disease heavily relies on imaging studies, which is a time consuming and a manual process performed by trained clinicians. Enhancing the accuracy and speed of the detection process can potentially have a significant impact on population health via early diagnosis and intervention. Besides the prevention, how to keep tracking the treatment effect for the patient with diabetic retinopathy is another crucial issue in personalized healthcare. Motivated by this, we propose a recognition framework, based on deep convolutional neural networks. Our recognition system predicts not only the severity levels of DR but also the location of symptoms at the pixel level. With the combination of DR severity levels and the segmented DR symptoms, our system can predict the severity levels of DR more accurately which could potentially provide another measurement to monitor the progression or regression of retinopathy with therapeutic intervention. For the classification of DR severity levels, the proposed lightweight network, DRNet-cla-v1 improves the classification performance in two aspects: (1) Without any fine-tuning, DRNet-cla-v1, combined with seven other boosting methods achieved 0.961 and 0.967 AUROC on the Messidor dataset for referable and non-referable screening, which outperforms state of the art (0.921 and 0.957). (2) Compared with CKML Net, VNXK, and Zoom-in-Net, DRNet-cla-v1 is more memory efficient with at least 5.23x fewer in total parameters and requires lower computation cost with at least 1.24x fewer in total FLOPs. For the segmentation of DR symptoms, the proposed network, DRNet-seg-v1, achieves 0.6894 average AUPRC on the IDRiD test dataset and outperforms the start of the art (0.6693). Finally, we use the linear SVM to fuse features extracted from the DRNet-cla-v1 and DRNet-seg-v1 and achieves the average accuracy 0.7281 on the IDRiD test dataset, which also outperforms the start of the art (0.6311).

摘要............................................. i
Abstract........................................ ii
誌謝............................................. iii
Contents........................................ iv
ListofFigures................................... vii
ListofTables ..................................... ix
Chapter1. Introduction .................................. 1
1.1 Background and Motivation for the Study .................... 1
1.2 Thesis Organization................................ 3
1.3 Diabetes Mellitus and Diabetic Retinopathy ................... 3
1.4 Deep Learning for Computer Vision and Medical Image Analysis....7
1.4.1 Deep Learning for Medical Image Analysis....7
1.4.2 Common Layers in Convolutional Neural Networks....8
1.4.3 Common Solvers in Convolutional Neural Networks.....12
Chapter2. Related Work ................................. 14
2.1 Overview ..................................... 14
2.2 Directed Approaches ............................... 14
2.3 Non-directed Approaches............................. 15
2.4 Fusion Approaches ................................ 15
Chapter3. TheProposedFramework........................... 16
3.1 Overview of the Proposed Framework ...................... 16
3.2 The Shared Pipeline................................ 16
3.2.1 Pre-ROIDetection ............................ 17
3.2.2 Post-ROIDetection............................ 18
3.2.3 Combination of the Pre-ROI and Post-ROI Detection................. 19
3.3 The Classification Pipeline ............................ 19
3.3.1 Image Enhancement ........................... 19
3.3.2 The Classification Network ....................... 21
3.3.3 Post-prediction .............................. 25
3.3.4 Eye blending ............................... 26
3.4 The Segmentation Pipeline ............................ 26
3.4.1 Image Enhancement ........................... 28
3.4.2 The Segmentation Network ....................... 29
3.5 TheFusion Pipeline................................ 37
3.5.1 Feature Selection............................. 38
3.5.2 The Linear Classifier........................... 39
Chapter4. Experiments.................................. 40
4.1 Fundus Image Datasets .............................. 40
4.1.1 Eye Picture Archive Communication System (EyePACS)
Dataset . . . 40
4.1.2 Messidor Dataset............................. 41
4.1.3 Indian Diabetic Retinopathy Image Dataset (IDRiD)....41
4.1.4 High-Resolution Fundus (HRF) ImageDataset...43
4.2 Evaluation Metrics ................................ 43
4.3 Experimental Results of the Shared Pipeline................... 44
4.4 Experimental Results of the Classification Pipeline.............. 45
4.4.1 Ablation Study on the EyePACSDataset..............46
4.4.2 Experiment Results on the Messidor Dataset.............. 47
4.4.3 Model Complexity Analysis ....................... 47
4.5 Experimental Results of the Segmentation Pipeline.............. 49
4.5.1 Ablation Study on the IDRiD ...................... 50
4.5.2 Comparison with Other Methods on the IDRiD.............. 51
4.6 Experimental Results of the Fusion Pipeline................... 51
4.6.1 Feature Selection of the DR classification on the IDRiD.............. 54
4.6.2 Feature Selection of the DME classification on the IDRiD.............. 54
4.6.3 Comparison with Other Methods on the IDRiD.............. 56
Chapter5. Conclusion................................... 57
5.1 Summary ..................................... 57
5.2 Future Work.................................... 58
References.......................................... 59

[1] IDF, “International diabetes federation. idf diabetes atlas, 8th edn.,” 2017.
[2] Danny Hope, “Photograph of the retina of the human eye, with overlay diagrams showing the positions and sizes of the macula, fovea, and optic disc,” https://en.wikipedia.org/ wiki/Fovea_centralis#/media/File:Macula.svg, 2014, Accessed: 2018-07-31.
[3] Prasanna Porwal, Samiksha Pachade, Ravi Kamble, Manesh Kokare, Girish Deshmukh, Vivek Sahasrabuddhe, Fabrice Meriaudeau, Gwenolé Quellec, Tom MacGillivray, Luca Giancardo, and Désiré Sidibé, “Diabetic retinopathy segmentation and grading challenge,” https://idrid.grand-challenge.org/home/, 2017, Accessed: 2018-07-31.
[4] Geert Litjens, Thijs Kooi, Babak Ehteshami Bejnordi, Arnaud Arindra Adiyoso Setio, Francesco Ciompi, Mohsen Ghafoorian, Jeroen A W M van der Laak, Bram Van Gin- neken, and Clara I Sánchez, “A survey on deep learning in medical image analysis,” Medical image analysis, vol. 42, pp. 60–88, 2017.
[5] Thijs Kooi, Geert Litjens, Bram van Ginneken, Albert Gubern-Mérida, Clara I Sánchez, Ritse Mann, Ard den Heeten, and Nico Karssemeijer, “Large scale deep learning for com- puter aided detection of mammographic lesions,” Medical image analysis, vol. 35, pp. 303–312, 2017.
[6] Mohsen Ghafoorian, Nico Karssemeijer, Tom Heskes, I W M van Uder, Frank-Erik de Leeuw, Elena Marchiori, Bram van Ginneken, and Bram Platel, “Non-uniform patch sampling with deep convolutional neural networks for white matter hyperintensity seg- mentation,” in Biomedical Imaging (ISBI), 2016 IEEE 13th International Symposium on. 2016, pp. 1414–1417, IEEE.
[7] Mark J J P van Grinsven, Bram van Ginneken, Carel B Hoyng, Thomas Theelen, and Clara I Sánchez, “Fast convolutional neural network training using selective data sam- pling: application to hemorrhage detection in color fundus images,” IEEE transactions on medical imaging, vol. 35, no. 5, pp. 1273–1284, 2016.
[8] AndreEsteva,BrettKuprel,RobertoANovoa,JustinKo,SusanMSwetter,HelenMBlau, and Sebastian Thrun, “Dermatologist-level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115, 2017.
[9] Wei Yang, Yingyin Chen, Yunbi Liu, Liming Zhong, Genggeng Qin, Zhentai Lu, Qianjin Feng, and Wufan Chen, “Cascade of multi-scale convolutional neural networks for bone suppression of chest radiographs in gradient domain,” Medical image analysis, vol. 35, pp. 421–433, 2017.
[10] Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov, “Dropout: a simple way to prevent neural networks from overfitting,” The Journal of Machine Learning Research, vol. 15, no. 1, pp. 1929–1958, 2014.
[11] Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous con- volution, and fully connected crfs,” arXiv preprint arXiv:1606.00915, 2016.
[12] Liang-Chieh Chen, George Papandreou, Florian Schroff, and Hartwig Adam, “Rethinking atrous convolution for semantic image segmentation,” arXiv preprint arXiv:1706.05587, 2017.
[13] Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam, “Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmenta- tion,” arXiv preprint arXiv:1802.02611, 2018.
[14] Gerald Liew, Michel Michaelides, and Catey Bunce, “A comparison of the causes of blindness certifications in England and Wales in working age adults (16-64 years), 1999- 2000 with 2009-2010,” BMJ open, vol. 4, no. 2, pp. e004015, 2014.
[15] Royal College of Ophthalmologists, “Diabetic retinopathy guide- lines,” https://www.rcophth.ac.uk/wp-content/uploads/2014/12/ 2013-SCI-301-FINAL-DR-GUIDELINES-DEC-2012-updated-July-2013.pdf, 2012, Accessed: 2018-07-31.
[16] ETDRS Levels, “INTERNATIONAL CLINICAL DIABETIC RETINOPATHY DIS- EASE SEVERITY SCALE DETAILED TABLE,” 2002.
[17] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun, “Delving Deep into Recti- fiers: Surpassing Human-Level Performance on ImageNet Classification,” CoRR, vol. abs/1502.0, 2015.
[18] “World health organization: Global report on diabetes,” http://apps.who.int/iris/ bitstream/10665/204871/1/9789241565257_eng.pdf?ua=1, Accessed: 2017-06-27.
[19] DonaldS.Fong,LloydAiello,ThomasW.Gardner,GeorgeL.King,GeorgeBlankenship, Jerry D. Cavallerano, Fredrick L. Ferris, and Ronald Klein, “Retinopathy in diabetes,” Diabetes Care, vol. 27, no. suppl 1, pp. s84–s87, 2003.
[20] Lisa Crossland, Deborah Askew, Robert Ware, Peter Cranstoun, Paul Mitchell, Andrew Bryett, and Claire Jackson, “Diabetic Retinopathy Screening and Monitoring of Early Stage Disease in Australian General Practice: Tackling Preventable Blindness within a Chronic Care Model,” J. Diabetes Res., vol. 2016, pp. 8405395, dec 2016.
[21] “Kaggle: Diabetic retinopathy detection,” https://www.kaggle.com/c/ diabetic-retinopathy-detectionua=1, 2015, Accessed: 2017-06-29.
[22] Jonathan Krause, Varun Gulshan, Ehsan Rahimy, Peter Karth, Kasumi Widner, Greg S Corrado, Lily Peng, and Dale R Webster, “Grader variability and the importance of ref- erence standards for evaluating machine learning models for diabetic retinopathy,” arXiv preprint arXiv:1710.01711, 2017.
[23] Shlomo Melmed, Williams textbook of endocrinology, Elsevier Health Sciences, 2016.
[24] Eric A Swanson, Charles P Lin, Joel S Schuman, William G Stinson, Warren Chang, and Kenton Gregory, “HHS Public Access,” vol. 254, no. 5035, pp. 1178–1181, 2015.
[25] MartiAHearst,SusanTDumais,EdgarOsuna,JohnPlatt,andBernhardScholkopf,“Sup- port vector machines,” IEEE Intelligent Systems and their applications, vol. 13, no. 4, pp. 18–28, 1998.
[26] Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton, “ImageNet Classification with Deep Convolutional Neural Networks,” in Adv. Neural Inf. Process. Syst. 25, F Pereira, C J C Burges, L Bottou, and K Q Weinberger, Eds., pp. 1097–1105. Curran Associates, Inc., 2012.
[27] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Fei-Fei Li, “ImageNet: A large-scale hierarchical image database.,” in CVPR. 2009, pp. 248–255, IEEE Computer Society.
[28] Jie Hu, Li Shen, and Gang Sun, “Squeeze-and-Excitation Networks,” arXiv Prepr. arXiv1709.01507, 2017.
[29] Forrest N Iandola, Matthew W Moskewicz, Khalid Ashraf, Song Han, William J Dally, and Kurt Keutzer, “SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and {<}1MB model size,” CoRR, vol. abs/1602.0, 2016.
[30] Ross Girshick, “Fast r-cnn,” in Proceedings of the IEEE international conference on computer vision, 2015, pp. 1440–1448.
[31] Joseph Redmon and Ali Farhadi, “{YOLO9000:} Better, Faster, Stronger,” CoRR, vol. abs/1612.0, 2016.
[32] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross B Girshick, “Mask r-cnn,” CoRR, vol. abs/1703.0, 2017.
[33] Ehsan Hosseini-Asl, Georgy Gimel’farb, and Ayman El-Baz, “Alzheimer’s Disease Di- agnostics by a Deeply Supervised Adaptable 3D Convolutional Network,” arXiv preprint arXiv:1607.00556, 2016.
[34] Jeremy Kawahara, Colin J Brown, Steven P Miller, Brian G Booth, Vann Chau, Ruth E Grunau, Jill G Zwicker, and Ghassan Hamarneh, “BrainNetCNN: convolutional neural networks for brain networks; towards predicting neurodevelopment,” NeuroImage, vol. 146, pp. 1038–1049, 2017.
[35] Olaf Ronneberger, Philipp Fischer, and Thomas Brox, “U-net: Convolutional networks for biomedical image segmentation,” in Int. Conf. Med. Image Comput. Comput. Interv. 2015, pp. 234–241, Springer.
[36] Fausto Milletari, Nassir Navab, and Seyed-Ahmad Ahmadi, “V-net: Fully convolutional neural networks for volumetric medical image segmentation,” in 3D Vision (3DV), 2016 Fourth International Conference on. 2016, pp. 565–571, IEEE.
[37] Mingchen Gao, Ziyue Xu, Le Lu, Aaron Wu, Isabella Nogues, Ronald M Summers, and Daniel J Mollura, “Segmentation label propagation using deep convolutional neural net- works and dense conditional random field,” in Biomedical Imaging (ISBI), 2016 IEEE 13th International Symposium on. 2016, pp. 1265–1268, IEEE.
[38] Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun, “Deep Residual Learning for Image Recognition,” CoRR, vol. abs/1512.0, 2015.
[39] Sergey Ioffe and Christian Szegedy, “Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift,” CoRR, vol. abs/1502.0, 2015.
[40] Matthew D Zeiler, “ADADELTA: an adaptive learning rate method,” arXiv preprint arXiv:1212.5701, 2012.
[41] John Duchi, Elad Hazan, and Yoram Singer, “Adaptive subgradient methods for online learning and stochastic optimization,” Journal of Machine Learning Research, vol. 12, no. Jul, pp. 2121–2159, 2011.
[42] Diederik P Kingma and Jimmy Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014.
[43] Tijmen Tieleman and Geoffrey Hinton, “Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude,” COURSERA: Neural networks for machine learning, vol. 4, no. 2, pp. 26–31, 2012.
[44] Yangqing Jia, Evan Shelhamer, Jeff Donahue, Sergey Karayev, Jonathan Long, Ross Gir- shick, Sergio Guadarrama, and Trevor Darrell, “Caffe: Convolutional architecture for fast feature embedding,” in Proc. 22nd ACM Int. Conf. Multimed. 2014, pp. 675–678, ACM.
[45] Balint Antal and Andras Hajdu, “An ensemble-based system for microaneurysm detection and diabetic retinopathy grading,” IEEE transactions on biomedical engineering, vol. 59, no. 6, pp. 1720, 2012.
[46] R Pires, S Avila, H F Jelinek, J Wainer, E Valle, and A Rocha, “Beyond Lesion-Based Diabetic Retinopathy: A Direct Approach for Referral,” 2017.
[47] H H Vo and A Verma, “New Deep Neural Nets for Fine-Grained Diabetic Retinopathy Recognition on Hybrid Color Space,” in 2016 IEEE Int. Symp. Multimed., 2016, pp. 209– 215.
[48] Zhe Wang, Yanxin Yin, Jianping Shi, Wei Fang, and Hongsheng Li, “Zoom-in-Net: Deep Mining Lesions for Diabetic Retinopathy Detection,” pp. 1–8, 2017.
[49] SajibKumarSaha,BasuraFernando,JorgeCuadros,DiXiao,andYogesanKanagasingam, “Deep Learning for Automated Quality Assessment of Color Fundus Images in Diabetic Retinopathy Screening,” arXiv preprint arXiv:1703.02511, 2017.
[50] Gwenolé Quellec, Katia Charrière, Yassine Boudi, Béatrice Cochener, and Mathieu Lamard, “Deep image mining for diabetic retinopathy screening,” Medical image analysis, vol. 39, pp. 178–193, 2017.
[51] Yehui Yang, Tao Li, Wensi Li, Haishan Wu, Wei Fan, and Wensheng Zhang, “Lesion detection and Grading of Diabetic Retinopathy via Two-stages Deep Convolutional Neural Networks,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. 2017, pp. 533–540, Springer.
[52] Varun Gulshan, Lily Peng, Marc Coram, Martin C. Stumpe, Derek Wu, Arunacha- lam Narayanaswamy, Subhashini Venugopalan, Kasumi Widner, Tom Madams, Jorge Cuadros, Ramasamy Kim, Rajiv Raman, Philip C. Nelson, Jessica L. Mega, and Dale R. Webster, “Development and Validation of a Deep Learning Algorithm for Detection of Di- abetic Retinopathy in Retinal Fundus Photographs.,” JAMA, vol. 304, no. 6, pp. 649–656, 2016.
[53] N Otsu, “A Threshold Selection Method from Gray-Level Histograms,” IEEE Trans. Syst. Man. Cybern., vol. 9, no. 1, pp. 62–66, 1979.
[54] MartinAFischlerandRobertCBolles,“Randomsampleconsensus:aparadigmformodel fitting with applications to image analysis and automated cartography,” Commun. ACM, vol. 24, no. 6, pp. 381–395, 1981.
[55] JohnCanny,“Acomputationalapproachtoedgedetection,”IEEETransactionsonpattern analysis and machine intelligence, , no. 6, pp. 679–698, 1986.
[56] A Polesel, G Ramponi, and V J Mathews, “Image enhancement via adaptive unsharp masking,” IEEE Trans. Image Process., vol. 9, no. 3, pp. 505–510, 2000.
[57] B. Graham, “Kaggle diabetic retinopathy detection competition report,” https://kaggle2.blob.core.windows.net/forum-message-attachments/88655/2795/ competitionreport.pdf, 2015, Accessed: 2017-07-06.
[58] Carlo Tomasi and Roberto Manduchi, “Bilateral filtering for gray and color images,” in Computer Vision, 1998. Sixth International Conference on. 1998, pp. 839–846, IEEE.
[59] Karen Simonyan and Andrew Zisserman, “Very Deep Convolutional Networks for Large- Scale Image Recognition,” CoRR, vol. abs/1409.1, 2014.
[60] MinLin,QiangChen,andShuichengYan,“NetworkInNetwork,”CoRR,vol.abs/1312.4, 2013.
[61] Sergey Ioffe and Christian Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” in International Conference on Machine Learning, 2015, pp. 448–456.
[62] Ling Li and Hsuan-Tien Lin, “Ordinal regression by extended binary classification,” in Advances in neural information processing systems, 2007, pp. 865–872.
[63] John A Nelder and Roger Mead, “A simplex method for function minimization,” Comput. J., vol. 7, no. 4, pp. 308–313, 1965.
[64] Harry Zhang, “The optimality of naive Bayes,” AA, vol. 1, no. 2, pp. 3, 2004.
[65] StephenMPizer,REugeneJohnston,JamesPEricksen,BonnieCYankaskas,andKeithE Muller, “Contrast-limited adaptive histogram equalization: speed and effectiveness,” in Visualization in Biomedical Computing, 1990., Proceedings of the First Conference on. 1990, pp. 337–345, IEEE.
[66] Wei Liu, Andrew Rabinovich, and Alexander C Berg, “Parsenet: Looking wider to see better,” arXiv preprint arXiv:1506.04579, 2015.
[67] Isabelle Guyon, Jason Weston, Stephen Barnhill, and Vladimir Vapnik, “Gene selection for cancer classification using support vector machines,” Machine learning, vol. 46, no. 1-3, pp. 389–422, 2002.
[68] Etienne Decencière, Xiwei Zhang, Guy Cazuguel, Bruno Lay, Béatrice Cochener, Caro- line Trone, Philippe Gain, Richard Ordonez, Pascale Massin, Ali Erginay, Béatrice Char- ton, and Jean-Claude Klein, “Feedback on a publicly distributed database: the Messidor database,” Image Anal. Stereol., vol. 33, no. 3, pp. 231–234, 2014.
[69] Deepali A. Godse and Dattatraya S. Bormane, “Automated localization of optic disc in retinal images,” 2013.
[70] Prasanna Porwal, Samiksha Pachade, Ravi Kamble, Manesh Kokare, Girish Deshmukh, Vivek Sahasrabuddhe, and Fabrice Meriaudeau, “Indian Diabetic Retinopathy Image Dataset (IDRiD): A Database for Diabetic Retinopathy Screening Research,” Data, vol. 3, no. 3, pp. 25, 2018.
[71] Attila Budai, Rüdiger Bock, Andreas Maier, Joachim Hornegger, and Georg Michelson, “Robust vessel segmentation in fundus images,” International journal of biomedical imag- ing, vol. 2013, 2013.
[72] J Richard Landis and Gary G Koch, “The measurement of observer agreement for cate- gorical data,” biometrics, pp. 159–174, 1977.
[73] Jacob Cohen, “Weighted kappa: Nominal scale agreement provision for scaled disagree- ment or partial credit.,” Psychological bulletin, vol. 70, no. 4, pp. 213, 1968.
[74] OpenCV Developers Team, “Open source computer vision library,” 2002.
[75] Eric Jones, Travis Oliphant, and Pearu Peterson, “{SciPy}: open source scientific tools for {Python},” 2014.
[76] Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, and Vincent Dubourg, “Scikit-learn: Machine learning in Python,” Journal of machine learning re- search, vol. 12, no. Oct, pp. 2825–2830, 2011.
[77] Clara I Sánchez, Meindert Niemeijer, Alina V Dumitrescu, Maria S A Suttorp-Schulten, Michael D Abramoff, and Bram van Ginneken, “Evaluation of a computer-aided diagnosis system for diabetic retinopathy screening on public data,” Invest. Ophthalmol. Vis. Sci., vol. 52, no. 7, pp. 4866–4871, 2011.
[78] Martín Abadi, Paul Barham, Jianmin Chen, Zhifeng Chen, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Geoffrey Irving, and Michael Isard, “Tensorflow: a system for large-scale machine learning.,” in OSDI, 2016, vol. 16, pp. 265–283.
[79] Ryan Poplin, Avinash V Varadarajan, Katy Blumer, Yun Liu, Michael V McConnell, Greg S Corrado, Lily Peng, and Dale R Webster, “Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning,” Nature Biomedical Engineering, vol. 2, no. 3, pp. 158, 2018.