This article presents a new algorithm for visual data augmentation based on statistical methods. The method includes an original approach to encoding contours as one-dimensional vectors, storing information about distances from the center of gravity to vertices at specific angles. An algorithm for generating new contours is proposed, based on the statistical characteristics of the original dataset and normal distribution. The key feature of the method is the preservation of important statistical properties of the original dataset, which is confirmed by mathematical proofs of two main statements about the invariance of mathematical expectation and variance. A visual example demonstrating the method's performance on a real contour is presented. The proposed approach has potential applications in various fields, including computer vision, medical imaging, and remote sensing, where generation and augmentation of object contour data play a crucial role. The method can be particularly useful in situations where collecting real data is difficult or resource-intensive. The main results were obtained through an analytical method – the developed mathematical model is supplemented by a random number generator from a distribution with parameters calculated based on the training dataset. The parameters are selected in such a way that the main statistical characteristics of the training dataset are preserved in the synthetic data, allowing for effective application of the proposed algorithm to a wide class of pattern recognition tasks.
1. Kalashnikov V.A. Study of augmentation methods in the problem of stone segmentation on a conveyor belt of a mining enterprise. Sovremennaya nauka: aktual'nye problemy teorii i praktiki. Seriya: Estestvennye i tekhnicheskie nauki = Modern Science: actual problems of theory and practice. Series: Natural and Technical Sciences. 2024;(1):69–71. (In Russ.).
2. Veselov D.I., Andriyanov N.A. Medical image segmentation using computer vision methods. In: Radiolokatsiya, navigatsiya, svyaz': Sbornik trudov XXX Mezhdunarodnoi nauchno-tekhnicheskoi konferentsii: Volume 2, 16–18 April 2024, Voronezh, Russia. Voronezh: Izdatel'skii dom VGU; 2024. pp. 75–80. (In Russ.).
3. Shelepov L.K., Polyakov A.N. Optimization of dataset creation for YOLO model training using remote sensing data. In: Far East Math – 2023: Materialy natsional'noi nauchnoi konferentsii, 04–09 December 2023, Khabarovsk, Russia. Khabarovsk: Pacific National University; 2024. pp. 146–150. (In Russ.).
4. Ahmad A., Andriyanov N.A., Soloviev V.I., Solomatin D.A. Application of deep learning for augmentation and generation of an underwater data set with industrial facilities. Vestnik Yuzhno-Ural'skogo gosudarstvennogo universiteta. Seriya: Komp'yuternye tekhnologii, upravlenie, radioelektronika = Bulletin of the South Ural State University. Series: Computer Technologies, Automatic Control, Radio Electronics. 2023;23(2):5–16. (In Russ.). https://doi.org/10.14529/ctcr230201
5. Trubin A.E., Morozov A.A., Zubanova A.E., Ozheredov V.A., Korepanova V.S. The method of preprocessing machine learning data for solving computer vision problems. Prikladnaya informatika = Journal of Applied Informatics. 2022;17(4):47–56. (In Russ.). https://doi.org/10.37791/2687-0649-2022-17-4-47-56
6. Gibadullin A.A. Generation of virtual reality. Akademicheskaya publitsistika. 2023;(12 2):231–233. (In Russ.).
7. Cheong Hou Y., Sahari K.S.M. Self-Generated Dataset for Category and Pose Estimation of Deformable Object. Journal of Robotics, Networking and Artificial Life. 2019;5(4):217–222. https://doi.org/10.2991/jrnal.k.190220.001
8. Gao X., Nguyen M., Yan W.Q. A High-Accuracy Deformable Model for Human Face Mask Detection. In: Image and Video Technology: 11th Pacific-Rim Symposium (PSIVT 2023): Proceedings, 22–24 November 2023, Auckland, New Zealand. Singapore: Springer; 2024. pp. 96–109. https://doi.org/10.1007/978-981-97-0376-0_8
9. Kramer D., Van der Merwe J., Lüthi M. A combined active shape and mean appearance model for the reconstruction of segmental bone loss. Medical Engineering & Physics. 2022;110. https://doi.org/10.1016/j.medengphy.2022.103841
10. Yuan H., Yanai K. Multi-Style Shape Matching GAN for Text Images. IEICE Transactions on Information and Systems. 2024;E107.D(4):505–514. https://doi.org/10.1587/transinf.2023IHP0010
11. Ribeiro T.F.R., Silva F., de C. Costa R.L. Reconstructing Spatiotemporal Data with C VAEs. In: Advances in Databases and Information Systems: 27th European Conference (ADBIS 2023): Proceedings, 04–07 September 2023, Barcelona, Spain. Cham: Springer; 2023. pp. 59–73. https://doi.org/10.1007/978-3-031-42914-9_5
Kalashnikov Vladimir Andreevich
ORCID | eLibrary |
Financial University under the Government of the Russian Federation
Moscow, Russia
