Classification of images based on the formation of independent cluster system within the structural descriptions of etalon dataset
Keywords:computer vision, structural image recognition, key point descriptor, BRISK detector, clustering, k-means method, cluster center, description relevance, software modeling, image database
AbstractThe subject of this article is the structural methods for image classification in the space of images as a set of descriptors of key points for recognizing visual objects in computer vision systems. The goal is the creation of an effective classification method based on the embedding of a system of independent clusters for the etalon dataset. Task: the development of classification models in the newly created space of images, analysis of their computational efficiency, the evaluation of classification efficiency with software modeling. The methods are: BRISK detector for generating key point descriptors, data mining, k-means method for data clustering, software modeling. The following results were obtained: models for classifying object descriptions based on a system of independent clusters and their centers are proposed that simplify data processing and increase implementation speed, a comparative analysis of the developed methods with known methods was performed. The software implementation of the embedded classification models has been performed, an experiment to explore their effectiveness and evaluate the processing time has been conducted. Conclusions. The contribution of the research is the development of an image classification method based on the implementation of a system of independent clusters for reference descriptions, which contributes to an in-depth data analysis. The method has been implemented in modifications of cluster representation matching and based on competitive analysis of descriptors. The practical importance of the work is the constructing of the classification models in the created data space, confirming the efficiency of the proposed modifications to data processing, developing software models for implementing methods in computer vision systems.
Gorokhovatskyi, V.A. (2003), Recognition of images in conditions of incomplete “information, KNURE, Kharkov, 112 p.
Xu, R. and Wunsch, D. (2009), Clustering, Wiley-IEEE Press, Hoboken, 358 p.
Aggarwa, C. and Reddy, C. (2014), Data Clustering. Algorithms and Application, CRC Press, Boca Raton, 652 p.
Gorokhovatskyi, V., Putyatin, E. and Stolyarov, V. (2017), “The study of the effectiveness of structural methods for classifying images using a cluster data model”, Radio Electronics, Computer Science, Control, No. 3 (42), pp. 78–85.
Gorokhovatskyi, V., Gadetska, S. and Ponomarenko, R. (2018), “Statistical distributions and chain representation of data when determining the relevance of structural descriptions of visual objects”, Control, navigation and communication systems, No. 6 (52), pp. 87–92.
Gorokhovatskyi O., Gorokhovatskyi V., and Peredrii, O. (2018), “Analysis of Application of Cluster Descriptions in Space of Characteristic Image Features”, Data, No. 3 (4), p. 52, DOI: http://doi.org/10.3390/data3040052
Litvin, V. and Krayovsky, V. (2009), “Modeling of rational agent activity based on adaptive technologies”, 4th International Scientific and Technical Conference on Computer Science and Information Technologies (CSIT), pp. 308-310.
Gorokhovatskyi, V. and Polyakova, T. (2018), The use of spatial structures of signs for the classification of images in comput-er vision, FLP Panov A.N., Kharkov, 120 p.
Eliseeva, I. and Rukavishnikov, V. (1977), Grouping, correlation, pattern recognition (Statistical methods for classifying and measuring relationships), Statistics, Moscow, 144 p.
Gorokhovatskyi, V.A. (2008), “Metrics on sets of key image points”, Bionics of intelligence, № 2 (69), pp. 45-50.
Kohonen, T. (2013), Self-organizing maps, BINOM. Laboratory of Knowledge, Moscow, 655 p.
Gorokhovatskyi, V. and Solodchenko, K. (2018), “The application system analysis and processing of the data bit in the methods of image classification for multiple key points”, Control, navigation and communication systems, No. 2 (48), pp. 63–67.
Medikovskiy, M., Tsmots, I., Tkachenko, R., Tsimbal, Yu. and Doroshenko, A. (2015), Intellectual components of integrated automation control systems, Vydavnytstvo Lʹvivsʹkoyi Politekhniky, Lviv, 280 p.
Leutenegger, S., Chli, M. and Siegwart, R. (2011), “BRISK: Binary Robust Invariant Scalable Keypoints”, 2011 International Conference on Computer Vision (ICCV), pp. 2548-2555.
Szeliski, R. (2010), Computer Vision: Algorithms and Applications, Springer, London, 979 p.
Gorokhovatskyi, V., Pupchenko, D. and Stiahlyk, N. (2020), “Research of data space transformation in Kohonen network training for methods of structural classification of images”, Advanced Information Systems, Vol. 4, No. 1, pp. 113-118, DOI: http://doi.org/10.20998/2522-9052.2020.1.17
Svyrydov, A., Kuchuk, H. and Tsiapa, O. (2018), “Improving efficienty of image recognition process: Approach and case study”, Proceedings in 2018 IEEE 9th International Conference on Dependable Systems, Services and Technologies (DES-SERT), pp. 593-597, DOI: http://dx.doi.org/10.1109/DESSERT.2018.8409201
Zhang, X., Felix, X. Yu, Chang, S.-F. and Karaman, S. (2017), “Learning Discriminative and Transformation Covariant Local Feature Detectors”, The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 6818-6826.