THE ALGORITHM OF DETERMINATION OF EYE FUNDUS VESSELS BLOOD FLOW CHARACTERISTICS ON VIDEOSEQUENCE
Abstract
The method of determination of the dynamic characteristics like the vessel diameter change, the linear and volume blood velocities in the vessels of the eye fundus is considered. Such characteristics allow to determine blood flow changes in the microvasculature affecting the blood flow in the brain, kidneys and coronary vessels. Developed algorithm includes four stages: the video sequence stabilization, the vessels segmentation with the help of a neural network, the determination of the instantaneous velocity in the vessels based on the optical flow and the analysis of the results.
About the Authors
O. V. Nedzvedz
Belarusian State University, Minsk;
The United Institute of Informatics Problems of the National Academy
of Sciences of Belarus, Minsk
Belarus
Senior Lecturer at the Department of Medical and Biological Physics
Belarus
Senior Lecturer at the Department of Medical and Biological Physics
S. V. Ablameyko
Belarusian State University, Minsk;
The United Institute of Informatics Problems of the National Academy
of Sciences of Belarus, Minsk
Belarus
D. Sc. (Technical Sciences), Academician of the National Academy of Science of Belarus, Professor at the Faculty of Mathematics and Mechanics
Belarus
D. Sc. (Technical Sciences), Academician of the National Academy of Science of Belarus, Professor at the Faculty of Mathematics and Mechanics
A. M. Nedzved
Belarusian State University, Minsk
Belarus
D. Sc. (Technical Sciences), Head of Department of Computer Technologies and Systems
Belarus
D. Sc. (Technical Sciences), Head of Department of Computer Technologies and Systems
A. V. Glinsky
Belarusian State Medical University Minsk
Belarus
Researcher at the Laboratory of Computing Technologies
Belarus
Researcher at the Laboratory of Computing Technologies
G. M. Karapetyan
Belarusian State Medical University Minsk
Belarus
Head of the Laboratory of Computing Technologies
Belarus
Head of the Laboratory of Computing Technologies
A. A. Anisimov
Belarusian State Medical University Minsk
Belarus
Assistant of the Department of Normal Physiology
Belarus
Assistant of the Department of Normal Physiology
I. B. Gurevich
Federal Research Center "Informatics and Management" of the Russian
Academy of Sciences (Dorodnitsyn Computing Center), Moscow
Russian Federation
Ph. D. (Physics and Mathematics), Head of Department of Mathematical Pattern Recognition and Methods of Combinatorial Analysis
Russian Federation
Ph. D. (Physics and Mathematics), Head of Department of Mathematical Pattern Recognition and Methods of Combinatorial Analysis
V. V. Yashina
Federal Research Center "Informatics and Management" of the Russian
Academy of Sciences (Dorodnitsyn Computing Center), Moscow
Russian Federation
Ph. D. (Physics and Mathematics), Leading Research Scientist at the Department of Mathematical Pattern Recognition and Methods of Combinatorial Analysis
Russian Federation
Ph. D. (Physics and Mathematics), Leading Research Scientist at the Department of Mathematical Pattern Recognition and Methods of Combinatorial Analysis
References
1. Algorithms for digital image processing in diabetic retinopathy / R. J. Winder [et al.] // Computerized Medical Imaging and Graphics. – 2009. – Vol. 33(8). – P. 608–622.
2. Felkel, P. Vessel tracking in peripheral CTA datasets – an overview / P. Felkel, R. Wegenkittl, A. Kanitsar // Computer Graphics (Spring Conference on). – Budmerice, Slovakia, 2001. – P. 232–239.
3. Buhler, K. Geometric methods for vessel visualization and quantification – a Survay / K. Buhler, P. Felkel, A. L. Cruz // Geometric Modelling for Scientific Visualization. – Berlin, Heidelberg: Springer, 2003. – P. 399–421.
4. Kirbas, C. A review of vessel extraction techniques and algorithms / C. Kirbas, F. Quek // ACM Computing. – 2004. – Vol. 36(2). – P. 81–121.
5. Mabrouk, M. S. Survey of retinal image segmentation and registration / M. S. Mabrouk, N. H. Solouma, Y. M. Kadah // International Journal on Graphics, Vision and Image Processing. – 2006. – Vol. 6(2). – P. 1–11.
6. Algorithms for the automated detection of diabetic retinopathy using digital fundus images : a review / O. Faust [et al.] // Journal of Medical Systems. – 2012. – Vol. 36(1). – P. 145–57.
7. Blood vessel segmentation methodologies in retinal images / M. M. Fraz [et al.] // Comput Methods Programs Biomed. – 2012. – Vol. 108(1). – P. 407–433.
8. Evaluation of a system for automatic detection of diabetic retinopathy from color fundus photographs in a large population of patients with diabetes / M. D. Abrаmoff [et al.] // Diabetes Care. – 2008. – Vol. 31(2). – P. 193–198.
9. Deep neural net-works segment neuronal membranes in electron microscopy images / D. C. Ciresan [et al.] // Advances in Neural Information Processing Systems 25 (NIPS 2012). – Harrahs and Harveys: Curran Associates, Inc., 2012. – P. 2852–2860.
10. Retinal vessel measurement: comparison between observer and computer driven methods / R. S. Newsom [et al.] // Graefes Arch. Clin. Exp. Ophthalmol. – 1992. – Vol. 230(3). – P. 221–225.
11. FlowNet: Learning Optical Flow with Convolutional Networks / A. Dosovitskiy [et al.] // Computer Vision (ICCV), IEEE Intern. Conf. – Chile, 2015. – P. 2758–2766.
12. Barron, J. L. Performance of optical flow techniques / J. L. Barron, D. J. Fleet, S. Beauchemin // International Journal of Computer Vision. – 1994. – Vol. 12(1). – P. 43–77.
13. Farnebäck, G. Two-Frame Motion Estimation Based on Polynomial Expansion / G. Farnebäck // Proceedings of the 13th Scandinavian Conf. on Image Analysis. – Halmstad, Sweden, 2003. – P. 363–370.
14. Detection of dynamical properties of flow in an eye vessels by video sequences analysis / A. Nedzved [et al.] // Intern. Conf. on Information and Digital Technologies. – Zilino, Slovakia, 2017. – P. 275–281.
Review
For citations:
Nedzvedz O.V., Ablameyko S.V., Nedzved A.M., Glinsky A.V., Karapetyan G.M., Anisimov A.A., Gurevich I.B., Yashina V.V. THE ALGORITHM OF DETERMINATION OF EYE FUNDUS VESSELS BLOOD FLOW CHARACTERISTICS ON VIDEOSEQUENCE. Informatics. 2018;15(1):92-102. (In Russ.)
Views: 832
Email this article 