Multiscale sequential convolutional neural networks for simultaneous detection of fovea and optic disc

Abstract Detecting the locations of the optic disc and fovea is a crucial task towards developing automatic diagnosis and screening tools for retinal disease. We propose to address this challenging problem by investigating the potential of applying deep learning techniques to this field. In the proposed method, simultaneous detection of the centers of the fovea and the optic disc (OD) from color fundus images is considered as a regression problem. A deep multiscale sequential convolutional neural network (CNN) is designed and trained. The publically available MESSIDOR and Kaggle datasets are used to train the network and evaluate its performance. The centers of the fovea and the OD in each image were marked by expert graders as the ground truth. The proposed method achieves an accuracy of 97%, 96.7% for the detection of the OD center and 96.6%, 95.6% for the detection of the foveal center of the MESSIDOR and Kaggle test sets respectively. Our promising results demonstrate the excellent performance of the proposed CNNs in simultaneously detecting the centers of both the fovea and OD without human intervention or handcrafted features. Moreover, we can localize the landmarks of an image in 0.007s. This approach could be used as a crucial part of automated diagnosis systems for better management of eye disease.

[1]  Geoffrey E. Hinton.  A Practical Guide to Training Restricted Boltzmann Machines , 2012, Neural Networks: Tricks of the Trade.

[2]  Heikki Kälviäinen,et al.  DIARETDB 0 : Evaluation Database and Methodology for Diabetic Retinopathy Algorithms , 2007 .

[3]  Tal Hassner,et al.  Age and gender classification using convolutional neural networks , 2015, 2015 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[4]  James E. Sheedy,et al.  Field Guide to Visual and Ophthalmic Optics , 2005 .

[5]  Joni-Kristian Kämäräinen,et al.  The DIARETDB1 Diabetic Retinopathy Database and Evaluation Protocol , 2007, BMVC.

[6]  Giorgio Terracina,et al.  Optic Disc Detection Using Fine Tuned Convolutional Neural Networks , 2016, 2016 12th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS).

[7]  Arturo Aquino,et al.  Establishing the macular grading grid by means of fovea centre detection using anatomical-based and visual-based features , 2014, Comput. Biol. Medicine.

[8]  Bram van Ginneken,et al.  Fast detection of the optic disc and fovea in color fundus photographs , 2009, Medical Image Anal..

[9]  J. Olson,et al.  Automatic detection of retinal anatomy to assist diabetic retinopathy screening , 2007, Physics in medicine and biology.

[10]  Marios S. Pattichis,et al.  Fast Localization and Segmentation of Optic Disk in Retinal Images Using Directional Matched Filtering and Level Sets , 2012, IEEE Transactions on Information Technology in Biomedicine.

[11]  András Hajdu,et al.  Detection of the optic disc in fundus images by combining probability models , 2015, Comput. Biol. Medicine.

[12]  Max A. Viergever,et al.  Ridge-based vessel segmentation in color images of the retina , 2004, IEEE Transactions on Medical Imaging.

[13]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[14]  Enrico Grisan,et al.  Detection of optic disc in retinal images by means of a geometrical model of vessel structure , 2004, IEEE Transactions on Medical Imaging.

[15]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[16]  Huiqi Li,et al.  Automated feature extraction in color retinal images by a model based approach , 2004, IEEE Transactions on Biomedical Engineering.

[17]  Grgoire Montavon,et al.  Neural Networks: Tricks of the Trade , 2012, Lecture Notes in Computer Science.

[18]  Amin Dehghani,et al.  Optic disc localization in retinal images using histogram matching , 2012, EURASIP Journal on Image and Video Processing.

[19]  Prabhat,et al.  Scalable Bayesian Optimization Using Deep Neural Networks , 2015, ICML.

[20]  Jacob Scharcanski,et al.  Fovea center detection based on the retina anatomy and mathematical morphology , 2011, Comput. Methods Programs Biomed..

[21]  Luca Maria Gambardella,et al.  Deep Neural Networks Segment Neuronal Membranes in Electron Microscopy Images , 2012, NIPS.

[22]  James T Schwiegerling,et al.  Field guide to visual and ophthalmic optics , 2004 .

[23]  Sushma G. Thorat.  Locating the Optic Nerve in a Retinal Image Using the Fuzzy Convergence of the Blood Vessels , 2014 .

[24]  Marc'Aurelio Ranzato,et al.  Unsupervised Learning of Invariant Feature Hierarchies with Applications to Object Recognition , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[25]  Antoni B. Chan,et al.  3D Human Pose Estimation from Monocular Images with Deep Convolutional Neural Network , 2014, ACCV.

[26]  Trevor Darrell,et al.  Do Convnets Learn Correspondence? , 2014, NIPS.

[27]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[28]  Rangaraj M. Rangayyan,et al.  Detection of the Optic Nerve Head in Fundus Images of the Retina Using the Hough Transform for Circles , 2010, Journal of Digital Imaging.

[29]  Karel J. Zuiderveld,et al.  Contrast Limited Adaptive Histogram Equalization , 1994, Graphics Gems.

[30]  Aliaa A. A. Youssif,et al.  Optic Disc Detection From Normalized Digital Fundus Images by Means of a Vessels' Direction Matched Filter , 2008, IEEE Transactions on Medical Imaging.

[31]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[32]  Stefan Holban,et al.  Segmentation of bone structure in X-ray images using convolutional neural network , 2013 .

[33]  Langis Gagnon,et al.  Procedure to detect anatomical structures in optical fundus images , 2001, SPIE Medical Imaging.

[34]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[35]  Bram van Ginneken,et al.  Segmentation of the Optic Disc, Macula and Vascular Arch in Fundus Photographs , 2007, IEEE Transactions on Medical Imaging.

[36]  Pablo Martínez-Cañada,et al.  Real-time tone mapping on GPU and FPGA , 2012, EURASIP J. Image Video Process..

[37]  J. van Leeuwen,et al.  Neural Networks: Tricks of the Trade , 2002, Lecture Notes in Computer Science.

[38]  Shijian Lu,et al.  Accurate and Efficient Optic Disc Detection and Segmentation by a Circular Transformation , 2011, IEEE Transactions on Medical Imaging.

[39]  José Manuel Bravo,et al.  Locating the fovea center position in digital fundus images using thresholding and feature extraction techniques , 2013, Comput. Medical Imaging Graph..

[40]  Kenneth W. Tobin,et al.  Detection of Anatomic Structures in Human Retinal Imagery , 2007, IEEE Transactions on Medical Imaging.

[41]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[42]  H. Kälviäinen,et al.  DIARETDB 1 diabetic retinopathy database and evaluation protocol , 2007 .

[43]  Laude,et al.  FEEDBACK ON A PUBLICLY DISTRIBUTED IMAGE DATABASE: THE MESSIDOR DATABASE , 2014 .

[44]  A.D. Hoover,et al.  Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response , 2000, IEEE Transactions on Medical Imaging.

[45]  András Hajdu,et al.  Combining algorithms for automatic detection of optic disc and macula in fundus images , 2012, Comput. Vis. Image Underst..

[46]  Andrea Giachetti,et al.  The use of radial symmetry to localize retinal landmarks , 2013, Comput. Medical Imaging Graph..

[47]  Frans Coenen,et al.  Automated "disease/no disease" grading of age-related macular degeneration by an image mining approach. , 2012, Investigative ophthalmology & visual science.

[48]  Hans Walther Larsen.  The ocular fundus: A color atlas , 1976 .