Comparing humans and deep learning performance for grading AMD: A study in using universal deep features and transfer learning for automated AMD analysis

Computers in Biology and Medicine

Volumn 82, Issue , 2017, Pages 80-86

  Burlina, Philippe ^a,b,c   Pacheco, Katia D ^d   Joshi, Neil ^a   Freund, David E ^a   Bressler, Neil M ^b  

^a JOHNS HOPKINS UNIVERSITY APPLIED PHYSICS LABORATORY   (United States)

^b JOHNS HOPKINS UNIVERSITY   (United States)

^c JOHNS HOPKINS UNIVERSITY   (United States)

^d Brazilian Center of Vision Eye Hospital   (Brazil)

Author keywords

Age related macular degeneration, (AMD); Deep Convolutional Neural Networks, (DCNNs); Deep learning; Retinal image analysis; Transfer learning; Universal features

Indexed keywords

ALDEHYDES; CLASSIFICATION (OF INFORMATION); CONVOLUTION; DEEP NEURAL NETWORKS; DIETARY SUPPLEMENTS; GRADING; IMAGE ANALYSIS; LEARNING SYSTEMS; NEURAL NETWORKS; OPHTHALMOLOGY; VISION;

AGE-RELATED MACULAR DEGENERATION; CONVOLUTIONAL NEURAL NETWORK; RETINAL IMAGE ANALYSIS; TRANSFER LEARNING; UNIVERSAL FEATURES;

DEEP LEARNING;

ACCURACY; AGE RELATED MACULAR DEGENERATION; ARTICLE; ARTIFICIAL NEURAL NETWORK; AUTOMATED RETINAL IMAGE ANALYSIS; AUTOMATION; CLASSIFICATION ALGORITHM; EYE FUNDUS; HUMAN; IMAGE ANALYSIS; MACHINE LEARNING; PRIORITY JOURNAL; RETINA IMAGE; SCORING SYSTEM; SUPPORT VECTOR MACHINE; ALGORITHM; AUTOMATED PATTERN RECOGNITION; CLASSIFICATION; COMPARATIVE STUDY; COMPUTER ASSISTED DIAGNOSIS; DIAGNOSTIC IMAGING; EARLY DIAGNOSIS; EVALUATION STUDY; FLUORESCENCE ANGIOGRAPHY; MACULAR DEGENERATION; OBSERVER VARIATION; PATHOLOGY; PROCEDURES; REPRODUCIBILITY; SENSITIVITY AND SPECIFICITY; SEVERITY OF ILLNESS INDEX;

ALGORITHMS; EARLY DIAGNOSIS; FLUORESCEIN ANGIOGRAPHY; HUMANS; IMAGE INTERPRETATION, COMPUTER-ASSISTED; MACHINE LEARNING; MACULAR DEGENERATION; OBSERVER VARIATION; PATTERN RECOGNITION, AUTOMATED; REPRODUCIBILITY OF RESULTS; SENSITIVITY AND SPECIFICITY; SEVERITY OF ILLNESS INDEX;

EID: 85012245192     PISSN: 00104825     EISSN: 18790534     Source Type: Journal    
DOI: 10.1016/j.compbiomed.2017.01.018     Document Type: Article

References (37)

1. [1] Subfoveal neovascular lesions in age-related macular degeneration: guidelines for evaluation and treatment in the macular photocoagulation study, Arch. Ophthalmol., 1099, pp. 1242–1257. URL +〈 http://dx.doi.org/10.1001/archopht.1991.01080090066027〉, 1991.
2. [2] Potential public health impact of age-related eye disease study results: AREDS report no. 11, Arch. Ophthalmol., 12111, pp. 1621–1624. URL+〈 http://dx.doi.org/10.1001/archopht.121.11.1621〉, 2003.
3. [3] Abramoff, M., Garvin, M., Sonka, M., Retinal imaging and image analysis. IEEE Rev. Biomed. Eng. 3 (2010), 169–208.
4. [4] Age-Related Eye Disease Study Research Group and others, The age-related eye disease study system for classifying age-related macular degeneration from stereoscopic color fundus photographs: the age-related eye disease study report number 6, Am. J. Ophthalmol., 2001, 1325, p. 668.
5. [5] Bird, A., Bressler, N., Bressler, S., Chisholm, I., Coscas, G., Davis, M., de Jong, P., Klaver, C., Klein, B., Klein, R., Mitchell, P., Sarks, J., Sarks, S., Soubrane, G., Taylor, H., Vingerling, J., An international classification and grading system for age-related maculopathy and age-related macular degeneration. Surv. Ophthalmol. 395 (1995), 367–374 http://www.sciencedirect.com/science/article/pii/S003962570580092X.
6. [6] K. Blinder, S. Bradley, N. Bressler, S. Bressler, G. Donati, Y. Hao, C. Ma, U. Menchini, J. Miller, M. Potter, C. Pournaras, A. Reaves, P. Rosenfeld, H. Strong, M. Stur, X. Su, G. Virgili, Treatment of Age-related Macular Degeneration with Photodynamic Therapy study group, Verteporfin in Photodynamic Therapy study group, Effect of lesion size, visual acuity, and lesion composition on visual acuity change with and without verteporfin therapy for choroidal neovascularization secondary to age-related macular degeneration: Tap and vip report no. 1, Am. J. Ophthalmol., 1363, pp. 407–418. URL 〈 http://dx.doi.org/10.1016/S0002-9394(03)00223-X〉, 2003.
7. [7] Bressler, N., Age-related macular degeneration is the leading cause of blindness. J. Am. Med. Assoc. 19115 (2004), 1900–1901 http://dx.doi.org/10.1001/jama.291.15.190.
8. [8] Bressler, N.M., Chang, T.S., Sun˜er, I.J., Fine, J.T., Dolan, C.M., Ward, J., Ianchulev, T., Vision- related function after ranibizumab treatment by better- or worse-seeing eye. Ophthalmology 1174 (2010), 747–756 e4 http://www.aaojournal.org/article/S0161-6420(09)00981-6/abstract.
9. [9] Burlina, P., Freund, D.E., Dupas, B., Bressler, N., Automatic screening of age-related macular degeneration and retinal abnormalities. IEEE Eng. Med. Biol., 2011, 3962–3966.
10. [10] P. Burlina, D.E. Freund, N. Joshi, Y. Wolfson, N. Bressler, Detection of age-related macular degeneration via deep learning, in: Proceedings of the International Symposium on Biomedical Imaging, 7493240, June 2016, pp. 184–188.
11. [11] G. Cardillo, Cohens kappa: compute the Cohen's kappa ratio on a 2×2 matrix. % 〈 http://www.mathworks.com/matlabcentral/fileexchange/15365〉, 2007.
12. [12] Dalal, N., Triggs, B., Histograms of oriented gradients for human detection. 2005 IEEE Comput. Vis. Pattern Recognit., 2005, 886–893.
13. [13] Fan, R.E., Chang, K.J., Hsieh, C.J., Wang, X.R., Lin, C.J., LIBLINEAR: A library for large linear classification. J. Mach. Learn. Res., 2008, 1871–1874.
14. [14] Feeny, A.K., Tadarati, M., Freund, D.E., Bressler, N.M., Burlina, P., Automated segmentation of geographic atrophy of the retinal epithelium via random forests in AREDS color fundus images. Comput Biol. Med. 65 (2015), 124–136.
15. [15] D.E. Freund, N. Bressler, P. Burlina, Automated detection of drusen in the macula, in: Proceedings of the IEEE International Symposium on Biomedical Imaging, 2009, pp. 61–64.
16. [16] R. Girshick, J. Donahue, T. Darrell, J. Malik, Rich feature hierarchies for accurate object detection and semantic segmentation, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2014.
17. [17] Holz, F.G., Strauss, E.C., Schmitz-Valckenberg, S., van Lookeren, Campagne, M., Geographic atrophy: clinical features and potential therapeutic approaches. Ophthalmology 1215 (2014), 1079–1091 http://www.sciencedirect.com/science/article/pii/S0161642013011020.
18. [18] Krizhevsky, A., Sutskever, I., Hinton, G., Imagenet classification with deep convolutional neural networks. Adv. Neural Inf. Process. Syst., 2012.
19. [19] Landis, J.R., Koch, G.G., The measurement of observer agreement for categorical data. Biometrics 33 (1977), 159–174.
20. [20] Lim, L.S., Mitchell, P., Seddon, J.M., Holz, F.G., Wong, T.Y., Age-related macular degeneration. Lancet 3799827 (2012), 1728–1738 http://www.sciencedirect.com/science/article/pii/S0140673612602827.
21. [21] D.G. Lowe, Distinctive Image Features from scale invariants keypoints, in: Proceedings of the International Journal of Computer Vision, 60(2), 2004, pp. 91–110.
22. [22] Rajagopalas, A.N., Burlina, P., Chellappa, R., Detection of people in images. Int. Jt. Conf. Neural Netw. 4 (1999), 2747–2752.
23. [23] A. Razavian, et al. CNN features off-the-shelf: an astounding baseline for recognition, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2014.
24. [24] J. Redmon, et al., You only look once: unified, real-time object detection, arXiv preprint arXiv, 2015, 1506.02640.
25. [25] S. Ren, K. He, R. Girshick, J. Sun, Faster R-CNN: Towards real-time object detection with region proposal networks, In Advances in Neural Information Processing, 2015, pp. 91–99.
26. [26] Russakovsky, O., et al. Imagenet large scale visual recognition challenge. Int. J. Comput. Vis. 115 (2015), 211–252.
27. [27] Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., LeCun, Y., Overfeat: integrated recognition, localization and detection using convolutional networks. arXiv Prepr. arXiv, 1312, 2013, 6229.
28. [28] K. Simonyan, Z. Zisserman, Very deep convolutional networks for large-scale image recognition, arXiv preprint, 2014, 1409.1556.
29. [29] C. Szegedy et al., Going deeper with convolutions, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015.
30. [30] The AREDS Research Group, Change in area of geographic atrophy in the age-related eye disease study: AREDS report number 26, Arch. Ophthalmol., 1279, pp. 1168–1174. URL +〈 http://dx.doi.org/10.1001/archophthalmol.2009.198〉, 2009.
31. [31] Tolentino, M.J., Dennrick, A., John, E., Tolentino, M.S., Drugs in Phase II clinical trials for the treatment of age-related macular degeneration. Expert Opin. Investig. Drugs 242 (2015), 183–199 http://dx.doi.org/10.1517/13543784.2015.961601.
32. [32] Trucco, E., Ruggeri, A., Karnowski, T., Giancardo, L., Chaum, E., Hubschman, J.P., et al. Validating retinal fundus image analysis algorithms: issues and a proposal. Investig. Ophthalmol. Vis. Sci. 545 (2013), 3546–3559.
33. [33] U.S. Department of Commerce USCB The 2012 Statistical Abstract The National Data Book, 2012.
34. [34] Vapnik, Statistical Learning Theory. 1998, Wiley, New York, 416–417.
35. [35] Age-Related Eye Disease Study (AREDS) Research Group, A randomized, placebo-controlled, clinical trial of high-dose supplementation with vitamins C and E, beta carotene, and zinc for age-related macular degeneration and vision loss: AREDS report no. 8, Arch. Ophthalmol., 119, 2001, pp. 1417–1436.
36. [36] 〈 https://github.com/sermanet/OverFeat〉.
37. [37] Pacheco, K., et al. Evaluation of automated drusen detection system for fundus photographs of patients with age-related macular degeneration. Invest. Ophthalmol. Vis. Sci., 2016 (1611-1611).