access icon free Automated diabetic retinopathy grading and lesion detection based on the modified R-FCN object-detection algorithm

In this work, we develop a computer-aided retinal image screening system that can perform automated diabetic retinopathy (DR) grading and DR lesion detection in retinal fundus images. We propose a modified object-detection method for this task via a region-based fully convolutional network (R-FCN). A feature pyramid network and a modified region proposal network are applied to enhance the detection of small objects. The DR-grading model based on the modified R-FCN is evaluated on the Messidor data set and images provided by the Shanghai Eye Hospital. High sensitivity of 99.39% and specificity of 99.93% are obtained on the hospital data. Moreover, high sensitivity of 92.59% and specificity of 96.20% are obtained on the Messidor data set. The modified R-FCN lesion-detection model is validated on the hospital data set and achieves a 92.15% mean average precision. The proposed R-FCN can efficiently accomplish DR grading and lesion detection with high accuracy.

Inspec keywords: diseases; medical image processing; blood vessels; image classification; biomedical optical imaging; feature extraction; object detection; eye; image segmentation

Other keywords: modified object-detection method; feature pyramid network; modified R-FCN object-detection algorithm; computer-aided retinal image screening system; Shanghai Eye Hospital; modified R-FCN lesion-detection model; Messidor data; hospital data; retinal fundus images; DR-grading model; region-based fully convolutional network; modified region proposal network; lesion detection

Subjects: Optical and laser radiation (medical uses); Computer vision and image processing techniques; Biology and medical computing; Patient diagnostic methods and instrumentation; Biomedical measurement and imaging; Optical, image and video signal processing

References

    1. 1)
      • 1. Williams, R.: ‘Epidemiology of diabetic retinopathy and macular oedema: a systematic review’, Eye, 2000, 18, (2), pp. 963983.
    2. 2)
      • 25. Shelhamer, E., Long, J., Darrell, T.: ‘Fully convolutional networks for semantic segmentation’, IEEE Trans. Pattern Anal. Mach. Intell., 2017, 39, (4), pp. 640651.
    3. 3)
      • 6. Gulshan, V., Peng, L., Coram, M., et al: ‘Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs’, J. Am. Med. Assoc., 2016, 316, (22), pp. 24022410.
    4. 4)
      • 8. Orlando, J. I., Prokofyeva, E., Fresno, M., et al: ‘An ensemble deep learning based approach for red lesion detection in fundus images’, Comput. Methods Programs Biomed., 2017, 153, pp. 115127.
    5. 5)
      • 2. Abràmoff, M., Lou, Y., Erginay, A., et al: ‘Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning’, Invest. Ophthalmol. Visual Sci., 2016, 57, (13), pp. 52005206.
    6. 6)
      • 3. Acharya, U.R., Ng, E.Y., Tan, J.H., et al: ‘An integrated index for the identification of diabetic retinopathy stages using texture parameters’, J. Med. Syst., 2011, 36, (3), pp. 20112020.
    7. 7)
      • 11. Zhang, H., Kyaw, Z., Yu, J., et al: ‘PPR-FCN: weakly supervised visual relation detection via parallel pairwise R-FCN’. Int. Conf. Computer Vision, Venice, Italy, 2017, pp. 42434251.
    8. 8)
      • 21. Uijlings, J., Sande, K., Gevers, T., et al: ‘Selective search for object recognition’, Int. J. Comput. Vis., 2013, 104, (2), pp. 154171.
    9. 9)
      • 30. Rekhi, R. S., Issac, A., Dutta, M. K., et al: ‘Automated classification of exudates from digital fundus images’. Int. Conf. and Workshop on Bioinspired Intelligence, Portugal, 2017, pp. 16.
    10. 10)
      • 5. Mookiah, M., Acharya, U., Chua, C., et al: ‘Computer-aided diagnosis of diabetic retinopathy: a review’, Comput. Biol. Med., 2013, 43, (12), pp. 21362155.
    11. 11)
      • 20. Girshick, R., Donahue, J., Darrell, T., et al: ‘Region-based convolutional networks for accurate object detection and segmentation’, IEEE Trans. Pattern Anal. Mach. Intell., 2015, 38, (1), pp. 142158.
    12. 12)
      • 15. Lin, T. Y., Dollár, P., Girshick, R., et al: ‘Feature pyramid networks for object detection’. Int. Conf. on Computer Vision and Pattern Recognition, Honolulu, Hawaii, 2017, pp. 936944.
    13. 13)
      • 28. Russakovsky, O., Deng, J., Su, H., et al: ‘Imagenet large scale visual recognition challenge’, Int. J. Comput. Vis., 2015, 115, (3), pp. 211252.
    14. 14)
      • 19. Girshick, R., Donahue, J., Darrell, T., et al: ‘Rich feature hierarchies for accurate object detection and semantic segmentation’. Int. Conf. on Computer Vision and Pattern Recognition, Columbus, OH, US, 2014, pp. 580587.
    15. 15)
      • 22. Krizhevsky, A., Sutskever, I., Hinton, G. E.: ‘Imagenet classification with deep convolutional neural networks’. Int. Conf. Proc. Advances in Neural Information Processing Systems, Lake Tahoe, NV, US, 2012, pp. 11061114.
    16. 16)
      • 9. Simonyan, K., Zisserman, A.: ‘Very deep convolutional networks for large-scale image recognition’. Int. Conf. on Learning Representations, San Diego, CA, US, 2015.
    17. 17)
      • 4. Agurto, C., Murray, V., Barriga, E., et al: ‘Multiscale AM-FM methods for diabetic retinopathy lesion detection’, IEEE Trans. Med. Imaging, 2010, 29, (2), pp. 502512.
    18. 18)
      • 14. Sang, N., Ni, Z.: ‘Gesture recognition based on R-FCN in complex scenes’, Huazhong Keji Daxue Xuebao, 2017, 45, (10), pp. 5458.
    19. 19)
      • 10. Dai, J., Li, Y., He, K., et al: ‘R-FCN: object detection via region-based fully convolutional networks’. Neural Information Processing System, Barcelona, Spain, 2016.
    20. 20)
      • 13. Rasmussen, C. B., Nasrollahi, K., Moeslund, T. B.: ‘R-FCN object detection ensemble based on object resolution and image quality’. Int. Joint Conf. on Computational Intelligence, Melbourne, Australia, 2017, pp. 110120.
    21. 21)
      • 12. Liu, L., Jiang, F., Shen, R.: ‘Highly occluded face detection: an improved R-FCN approach’. Int. Conf. Neural Information Processing, Guangzhou, China, 2017, pp. 592601.
    22. 22)
      • 26. He, K., Zhang, X., Ren, S., et al: ‘Deep residual learning for image recognition’. Int. Conf. Computer Vision and Pattern Recognition, Las Vegas, NV, US, 2016, pp. 770778.
    23. 23)
      • 27. Neubeck, A., Gool, L.: ‘Efficient non-maximum suppression’. Int. Conf. Pattern Recognition, Hong Kong, China, 2006, pp. 850855.
    24. 24)
      • 29. Everingham, M., Gool, L., Williams, C. K., et al: ‘The pascal visual object classes (VOC) challenge’, Int. J. Comput. Vis., 2010, 88, (2), pp. 303338.
    25. 25)
      • 17. Wang, S., Tang, H. L., Turk, L. I. A., et al: ‘Localizing microaneurysms in fundus images through singular spectrum analysis’, IEEE Trans. Biomed. Eng., 2017, 64, (5), pp. 9901002.
    26. 26)
      • 18. Szegedy, C., Ioffe, S., Vanhoucke, V., et al: ‘Inception-v4, inception-ResNet and the impact of residual connections on learning’, Pre-Print Version. Available at https://arxiv.org/pdf/1712.03400.pdf.
    27. 27)
      • 23. Girshick, R.: ‘Fast R-CNN’. Int. Conf. on Computer Vision, Santiago, Chile, 2015, pp. 14401448.
    28. 28)
      • 7. Pratt, H., Coenen, F., Broadbent, D. M., et al: ‘Convolutional neural networks for diabetic retinopathy’, Procedia Comput. Sci., 2016, 90, pp. 200205.
    29. 29)
      • 24. Ren, S., He, K., Girshick, R., et al: ‘Faster R-CNN: towards real-time object detection with region proposal networks’, IEEE Trans. Pattern Anal. Mach. Intell., 2017, 39, (6), pp. 11371149.
    30. 30)
      • 16. Antal, B., Hajdu, A.: ‘An ensemble-based system for microaneurysm detection and diabetic retinopathy grading’, IEEE Trans. Biomed. Eng., 2012, 59, (6), pp. 17201726.
http://iet.metastore.ingenta.com/content/journals/10.1049/iet-cvi.2018.5508
Loading

Related content

content/journals/10.1049/iet-cvi.2018.5508
pub_keyword,iet_inspecKeyword,pub_concept
6
6
Loading