INVESTIGATION OF THE EFFECTIVENESS OF DATA FEATURES DISTRIBUTION FOR IMAGE RELEVANCE ESTIMATION
DOI:
https://doi.org/10.26906/SUNZ.2020.1.068Keywords:
computer vision, structural image classification methods, key point, ORB detector, descriptor, fragment data distribution, descriptive relevance, Manhattan metric, speed of relevance estimationAbstract
The subjects of the paper are the models of data attributes distribution of key point descriptors for recognition and classification of visual objects in computer vision systems. The goal is the investigation of the modification of an image structural classification method based on the matching fragment distributions of image descriptor set. The tasks include the development of mathematical and software models of efficient relevance estimation based on the data distribution, investigation of the properties of these models, evaluation of the effectiveness of image classification. Methods below are used: an ORB detector to form the key point descriptors, data mining, methods for construction of the bitwise data distribution, a method of metric relevance estimation, software modeling. The following results were obtained. The transition from the sets of descriptors to distributions of fragments, the construction and matching of distributions provide the decent classification performance. Classification is performed several times faster compared to the usage of descriptor sets directly. Conclusion. The contribution of the paper is the improvement of the structural image classification method with the description of a block structure using distribution values for fragments of the set of key point descriptors. The practical significance of the paper is the increase of image relevance calculation speed, verification of the effectiveness of the proposed attribute space with image examples, obtaining of an application software models for research and implementation of classification methods in computer vision systemsDownloads
References
Гороховатский В.А. (2014) Структурный анализ и интеллектуальная обработка данных в компьютерном зрении, Харьков, Компания СМИТ, 316 с.
Гороховатський В.О., Гадецька С.В., Пономаренко Р.П. (2018) Статистичні розподіли та ланцюжкове подання даних при визначенні релевантності структурних описів візуальних об’єктів. Системи управління, навігації та зв’язку, No6 (52), с. 87–92.
Gorokhovatsky, V.O. and Gadetska, S.V. (2019) Determination of Relevance of Visual Object Images by Application of Statistical Analysis of Regarding Fragment Representation of their Descriptions, Telecommunications and Radio Engineering, 78 (3), pp. 211–220. – doi: 10.1615/TelecomRadEng.v78.i3.20.
Гороховатский В.А., Ересько Ю.Н., Путятин Е.П., Стрельченко В.И. (1990) Локализация объектов на изображениях визуальных сцен. Автометрия, No6, с. 3-7.
Gorokhovatsky V.A., Gorokhovatsky A. V., Berestovsky A.Ye. (2016) Intellectual Data Processing and Self-Organization of Structural Features at Recognition of Visual Objects. Telecommunications and Radio Engineering, vol. 75, No 2. – P. 155–168.
Gorokhovatskiy V.A. (2011) Compression of Descriptions in the Structural Image Recognition. Telecommunications and Radio Engineering, vol. 70, No 15. – р. 1363–1371.
Гороховатський В.О., Руденко Д.О., Сірик Т.О. (2019) Дослідження системи ієрархічних ознак при блочному поданні опису у складі множини ключових точок зображення. Системи управління, навігації та зв’язку, No2 (54), с. 69–73. – doi: 10.26906/SUNZ.2019.1.003.
Rublee, E., Rabaud, V., Konolige, K., and Bradski, G., (2011) ORB: an efficient alternative to SIFT or SURF, IEEE International Conference on Computer Vision (ICCV), Proceedings, pp. 2564-2571.
Вадзинский Р. (2008) Статистические вычисления в среде EXCEL. Библиотека пользователя, СПб., 608 с.
Прохоренок Н.А. OpenCV и Java. (2018), Обработка изображений и компьютерное зрение, СПб., БХВ-Петербург, 320 с.
OpenCV Open Source Computer Vision [Електронний ресурс] – Режим доступу до ресурсу: https://docs.opencv.org/master/index.html.
Зображення нових монет Національного банку України [Електронний ресурс]. – Режим доступу до ресурсу: https://telegraf.design/news/natsionalnij-bank-prezentuvav-novi-obigovi-moneti.
Vinay, A., Cholin, A.S., Bhat, A.D., Murthy, K.B., Natarajan, S. (2018) An Efficient ORB based Face Recognition framework for Human-Robot Interaction. Procedia Computer Science 2018, 133, 913–923.
Y Zhang, C Li, C Cao et al. (2018) An Improved ORB Feature Point Matching Algorithm. Proceedings of the 2018 2nd International Conference on Computer Science and Artificial Intelligence, pp. 207-211.
Bian J, Lin W Y, Matsushita Y et al. (2017) GMS: Grid-Based Motion Statistics for Fast, Ultra-Robust Feature Correspondence. IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
Мариношенко А. П. (2018) Использование методов особых точек с целью улучшения идентификации и детектирования навигационных сигналов. Молодий вчений. — 2018. — No4.
Гороховатский В.А. (2008) Иерархия пространственных отношений структурных признаков в задачах сопоставления визуальных объектов. Системи управління, навігації та зв’язку: зб. наук. пр., 3(7), с. 85–89.
Porter, F.C. (2008) Testing Consistency of Two Histograms. ArXiv e-prints, pp. 1–35.
