P.O. Arkhipov1, S.L. Philippskih2, M.V. Tsukanov3

1–3 Branch Director, Orel Branch of the Federal Research Center “Computer Science and Control” of the RAS
(Orel, Russia)
1 arpaul@mail.ru, 2 philippsl@mail.ru, 3 tsukanov.m.v@yandex.ru

The article discusses a method for integrating a deep learning module into special software based on classical computer vision methods. This software implements the functions of creating panoramic images and their subsequent comparison with the ability to highlight different areas. The integration of deep learning methods is aimed at automating the classification of found objects and solving problems associated with identifying a large number of small objects and noise anomalies. In the process of carrying out the research, an additional neural network generative model was created that improves the quality of classification by increasing the size of the training dataset. This model was implemented in the form of an ensemble of variational autoencoders, which made it possible to simplify the learning process and fine-tuning of neural networks. Also, in order to assess the quality of the ensemble of variational autoencoders, a new method for checking the generated synthetic data was developed. The result of the integration was improved special software capable of classifying the detected found objects of interest with high accuracy.

Arkhipov P.O., Philippskih S.L., Tsukanov M.V. A New Method for Integrating a Deep Learning Module into Special Object Classification Software. Highly Available Systems. 2024. V. 20. № 1. P. 38−45. DOI: https://doi.org/ 10.18127/j20729472-202401-04 (in Russian)

1. Svidetel'stvo o gosudarstvennoj registracii programmy dlya EVM № 2022615139. Programma dlya EVM Programmnaya sistema obnaruzheniya anomalij na raznovremennyh panoramah obsleduemoj mestnosti (CL_22). P.O. Arhipov, A.V. Kolesnik, M.V. Cukanov. Zaregistrirovano v Reestre programm dlya EVM 30 marta 2022 g. 1 s. (in Russian).
2. Arhipov P.O., Cukanov M.V. Algoritmicheskaya model' obnaruzheniya anomalij na raznovremennyh panoramah. Sistemy vysokoj dostupnosti. 2021. T. 17. № 2. S. 5−10. DOI: 10.18127/j20729472-202102-01 (in Russian).
3. Arhipov P.O., Cukanov M.V., Trofimenkov A.K. Primenenie nechetkih mnozhestv v zadache obnaruzheniya anomalij na sravnivaemyh panoramah. Sistemy vysokoj dostupnosti. 2023. T. 19. № 2. S. 46−54. DOI: 10.18127/j20729472-202302-04 (in Russian).
4. Arkhipov P.O., Philippskih S.L. Building an ensemble of convolutional neural networks for classifying panoramic images. Pattern Recognition and Image Analysis. 2022. V. 32. № 3. P. 511–514. DOI: 10.1134/S1054661822030051.
5. Arhipov P.O., Filippskih S.L., Cukanov M.V. Razrabotka novoj modeli svertochnoj nejronnoj seti dlya klassifikacii anomalij na panoramah. Informatika i ee primeneniya. 2023. T. 17. Vyp. 1. S. 50–56. DOI: 10.14357/19922264230107 (in Russian).
6. Arhipov P.O., Filippskih S.L. Raspoznavanie anomalij na raznovremennyh panoramah s ispol'zovaniem nejrosetevogo pohoda konsolidacii modelej. Sistemy i sredstva informatiki. 2023. T. 33. № 2. S. 13–24. DOI:10.14357/08696527230202 (in Russian).
7. Svidetel'stvo o gosudarstvennoj registracii programmy dlya EVM № 2023685690. Programmnaya sistema opredeleniya i klassifikacii anomalij na sravnivaemyh panoramah, poluchennyh pri provedenii aerofotos"emki s BPLA. P.O. Arhipov, S.L. Filippskih, M.V. Cukanov. Zaregistrirovano v Reestre programm dlya EVM 29.11.2023 g. 1 s. (in Russian).
8. Zhu P., Wen L., Du D. et al. Detection and Tracking Meet Drones Challenge. Cornell University, 4 Oct 2021. arXiv: 2001.06303v3 [cs.CV].
9. Monarch R. Human-in-the-Loop Machine Learning. Active learning and annotation for human-centered AI. Manning Publications Co. 2021.
10. Kingma D., Welling M. Auto-Encoding Variational Bayes. Machine Learning Group Universiteit van Amsterdam, 10 Dec 2022. arXiv:1312.6114v11 [stat.ML].
11. Hou X. et al. Deep Feature Consistent Variational Autoencoder. University of Nottingham, 2 October 2016. arXiv:1610.00291v1 [cs.CV].
12. Foster D. Generative Deep Learning, 2nd Edition. O'Reilly Media, Inc., Sebastopol, CA (2023).
13. CKP «INFORMATIKA». URL: https://www.frccsc.ru/ckp (data obrashcheniya: 02.04.2024) (in Russian).
14. Lin T.-Y., Goyal P., et al. Focal Loss for Dense Object Detection, 7 Feb 2018. arXiv:1708.02002v2 [cs.CV].
15. Kullback S., Leibler R.A. On information and sufficiency. The Annals of Mathematical Statistics. 1951. V. 22. № 1. P. 79–86.