Adamov A.A., Gndoyan I.A., Dyatchina A.I., Khramov V.N. Development of a classifier of photo images of pathologies for an ultra-small data set
- Details
- Hits: 207
https://doi.org/10.15688/mpcm.jvolsu.2023.1.3
Anton A. Adamov
Candidate of Technical Sciences, Junior Researcher at the Institute of Mathematics and Information Technology, senior lecturer at the Departments of Radiophysics,
Volgograd State University
This email address is being protected from spambots. You need JavaScript enabled to view it.
,
This email address is being protected from spambots. You need JavaScript enabled to view it.
,
This email address is being protected from spambots. You need JavaScript enabled to view it.
https://orcid.org/0000-0002-7394-0744
Prosp. Universitetsky, 100, 400062 Volgograd, Russian Federation
Irina A. Gndoyan
Doctor of Medical Sciences, Head Of the Department of Ophthalmology,
Volgograd State Medical University
This email address is being protected from spambots. You need JavaScript enabled to view it.
https://orcid.org/0000-0001-7581-9473
pl. Pavshikh Bortsov 1, 400131 Volgograd, Russian Federation
Alena I. Dyatchina
Postgraduate student, Department of Ophthalmology,
Volgograd State Medical University
This email address is being protected from spambots. You need JavaScript enabled to view it.
https://orcid.org/0000-0001-9632-5800
pl. Pavshikh Bortsov 1, 400131 Volgograd, Russian Federation
Vladimir N. Khramov
Candidate of Physical and Mathematical Sciences, Associate Professor, Department of Radiophysics,
Volgograd State University
This email address is being protected from spambots. You need JavaScript enabled to view it.
,
This email address is being protected from spambots. You need JavaScript enabled to view it.
https://orcid.org/0000-0001-8988-0929
Prosp. Universitetsky, 100, 400062 Volgograd, Russian Federation
Abstract. The purpose of the work: to create an algorithm and implement it in a software tool for classifying photographic images of pathology of the central region of the human fundus, detected by autofluorescence research, according to 8 types-patterns: normal, minimal changes, focal, spotted, linear, lace-like, reticular, speckled. Methods: machine learning algorithms (convolutional neural networks) and computer vision (histogram methods, perceptual hash algorithms). The main feature of the task: an ultra-small set of unique photoimages with an accurately diagnosed type of pathology (18 pieces). The accuracy of forecasts when solving a problem using a neural network is 12.5%. The accuracy of the predictions of the developed algorithm using a combination of histograms, perceptual hash and 1 reference photo of the normal state of the fundus is 60% when selecting the classifier parameters from a set of 1 photo for 1 pathology. When using 3 reference photos, the norm is 85%. The proposed solution can be used in medicine, ophthalmology, photonics and optics of biological tissues, machine learning for both research and educational purposes.
Key words: photo image processing, computer vision, machine learning, image classification, histogram, perceptual hash, ophthalmological diagnostics, computerization of medicine.
Development of a classifier of photo images of pathologies for an ultra-small data set by Adamov A. A., Gndoyan I. A., Dyatchina A. I., Khramov V. N. is licensed under a Creative Commons Attribution 4.0 International License.
Citation in English: Mathematical Physics and Computer Simulation. Vol. 26 No. 1 2023, pp. 33-48