|
DOI: 10.14489/vkit.2026.02.pp.013-022
Наталенко Д. Н. ПОВТОРНАЯ ИДЕНТИФИКАЦИЯ ДИКИХ ЖИВОТНЫХ НА ИЗОБРАЖЕНИЯХ С ПРИМЕНЕНИЕМ МЕТОДОВ ГЛУБОКОГО ОБУЧЕНИЯ (c. 13-22)
Аннотация. Рассмотрена задача повторной идентификации диких животных на изображениях, полученных от фоторегистраторов, устанавливаемых в национальных парках и заповедниках. Предложен метод на основе сверточной нейронной сети с предварительной семантической сегментацией объекта интереса. Для повышения точности идентификации дополнительно используются временные метки и идентификаторы камер, что позволяет обеспечивать пространственно-временной контекст съемки и снижать вероятность ошибок. В качестве обучающего набора данных использовались изображения, собранные в природном парке «Ергаки» Красноярского края в 2012–2021 гг., а также изображения из открытых источников сети Интернет. Экспериментальные результаты, полученные для трех классов животных, демонстрируют высокую точность: 90,92 % для медведей, 83,94 % для волков и 90,60 % для оленей. Установлено, что использование семантической сегментации само по себе позволило повысить производительность примерно на 2 %. В сравнительных испытаниях предложенный подход превзошел другие методы на 4…8 % для медведей и оленей, а также показал конкурентоспособность в случае с волками.
Ключевые слова: повторная идентификация животных; фоторегистраторы; нейронные сети; временные метки; экологический мониторинг.
Natalenko D. N. RE-IDENTIFICATION OF WILD ANIMALS IN IMAGES USING DEEP LEARNING METHODS (pp. 13-22)
Abstract. This paper addresses the significant challenge of automated wildlife re-identification in images captured by camera traps in national parks and nature reserves. A novel deep learning-based framework is proposed, which synergistically combines preliminary semantic segmentation for precise animal isolation with a modified ResNet-50 architecture for robust, discriminative feature extraction. A key innovation is the strategic incorporation of spatiotemporal context through camera IDs and timestamps, which substantially enhances accuracy by filtering implausible matches and reducing false positives. The model was trained on a comprehensive, multi-source dataset from the Ergaki Nature Park (2012–2021) and public internet repositories, augmented to improve generalization across challenging environmental conditions. Rigorous evaluation on three animal classes (bears, wolves, deer) demonstrates the method's high efficacy, achieving mean accuracy rates at 90.92 %, 83.94 %, and 90.60 %, respectively. The integration of semantic segmentation alone contributed to a consistent ~2 % performance gain. In a comparative study, our approach outperformed contemporary state-of-the-art methods (PPGNet R-50 and PGCFL) by 4–8 % for bears and deer, while remaining competitive for wolves. These results validate the method's superior capability for reliable, automated ecological monitoring and scalable population analysis, with potential applications in biodiversity conservation and wildlife management.
Keywords: Animal re-identification; Camera traps; Neural networks; Timestamps; Wildlife monitoring.
Д. Н. Наталенко (Сибирский государственный университет науки и технологий имени академика М. Ф. Решетнева, Красноярск, Россия) E-mail:
Этот e-mail адрес защищен от спам-ботов, для его просмотра у Вас должен быть включен Javascript
D. N. Natalenko (Reshetnev Siberian State University of Science and Technology, Krasnoyarsk, Russia)
1. Deep Residual Learning for Image Recognition / K. He, X. Zhang, S. Ren et al. // IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, Nevada, USA. 27–30 June 2016. 2016. P. 770–778. DOI: 10.1109/CVPR.2016.90 2. Revisiting animal photo-identification using deep metric learning and network analysis / V. Miele, G. Dussert, B. Spataro et al. // Methods in Ecology and Evolution. 2021. V. 12, Is. 5. P. 863–873. 3. Ravoor P. C., Tsb S. Deep Learning Methods for Multi-Species Animal Re-Identification and Tracking – a Survey // Computer Science Review. 2020. V. 38, Is. 6. Art. 100289. 4. Bąk S., Carr P. One-Shot Metric Learning for Person Re-Identification // IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, Hawaii, USA. 21–26 July 2017. P. 2990–2999. DOI: 10.1109/CVPR.2017.171 5. Addressing the Elephant in the Room: Robust Animal Re-Identification with Unsupervised Part-Based Feature Alignment / Y. Yu, V. Vidit, A. Davydov et al. // CVPR Workshop on Computer Vision for Animals (CV4Animals). Seattle, WA, USA. 17 June 2024. Art. 13781. 6. Person Transfer GAN to Bridge Domain Gap for Person Re-Identification / L. Wei, S. Zhang, W. Gao et al. // IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. 18–23 June 2018. P. 79–88. 7. Joint Detection and Identification Feature Learning for Person Search / T. Xiao, S. Li, B. Wang et al. // IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA, 21–26 Jule 2017. P. 3415–3424. DOI: 10.1109/CVPR.2017.360 8. Scalable Person Re-Identification: A Benchmark / L. Zheng, L. Shen, L. Tian et al. // IEEE International Conference on Computer Vision (ICCV). Santiago, Chile, 7–13 December 2015. P. 1116–1124. DOI: 10.1109/ICCV.2015.133 9. Gamani A. R. A., Arhin I., Asamoah A. K. Performance Evaluation of YOLOv8 Model Configurations, for Instance Segmentation of Strawberry Fruit Development Stages in an Open Field Environment // Adrena Kyeremateng Asamoah Faculty of Agriculture, KNUST. 2024. Art. 2408.05661. 10. Sapkota R., Ahmed D., Karkee M. Comparing YOLOv8 and Mask R-CNN for instance segmentation in complex orchard environments // Artificial Intelligence in Agriculture. 2024. V. 13. P. 84–99. 11. Liu C., Zhang R., Guo L. Part-Pose Guided Amur Tiger Re-Identification // IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). Seoul, Korea (South), 27–28 October 2019. Art. 00042. DOI: 10.1109/ICCVW.2019.00042 12. Pose-Guided Complementary Features Learning for Amur Tiger Re-Identification / N. Liu, Q. Zhao, N. Zhang et al. // IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). Seoul, Korea (South), 27–28 October 2019. Art. 00038. DOI: 10.1109/ICCVW.2019.00038
1. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770–778). https://doi.org/10.1109/CVPR.2016.90 2. Miele, V., Dussert, G., Spataro, B., Chamaillé-Jammes, S., & Benhamou, S. (2021). Revisiting animal photo-identification using deep metric learning and network analysis. Methods in Ecology and Evolution, 12(5), 863–873. https://doi.org/10.1111/2041-210X.13577 3. Ravoor, P. C., & TSB, S. (2020). Deep learning methods for multi-species animal re-identification and tracking – A survey. Computer Science Review, 38, Article 100289. https://doi.org/10.1016/j.cosrev.2020.100289 4. Bąk, S., & Carr, P. (2017). One-shot metric learning for person re-identification. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 2990–2999). https://doi.org/10.1109/CVPR.2017.171 5. Yu, Y., Vidit, V., Davydov, A., Brattoli, B., Ommer, B., & Weinzaepfel, P. (2024). Addressing the elephant in the room: Robust animal re-identification with unsupervised part-based feature alignment. In CVPR Workshop on Computer Vision for Animals (CV4Animals). Seattle, WA, USA. arXiv:2405.13781 6. Wei, L., Zhang, S., Gao, W., & Tian, Q. (2018). Person transfer GAN to bridge domain gap for person re-identification. In 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 79–88). https://doi.org/10.1109/CVPR.2018.00017 7. Xiao, T., Li, S., Wang, B., Lin, L., & Wang, X. (2017). Joint detection and identification feature learning for person search. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 3415–3424). https://doi.org/10.1109/CVPR.2017.360 8. Zheng, L., Shen, L., Tian, L., Wang, S., Wang, J., & Tian, Q. (2015). Scalable person re-identification: A benchmark. In 2015 IEEE International Conference on Computer Vision (ICCV) (pp. 1116–1124). https://doi.org/10.1109/ICCV.2015.133 9. Gamani, A. R. A., Arhin, I., & Asamoah, A. K. (2024). Performance evaluation of YOLOv8 model configurations, for instance segmentation of strawberry fruit development stages in an open field environment. arXiv. https://doi.org/10.48550/arXiv.2408.05661 10. Sapkota, R., Ahmed, D., & Karkee, M. (2024). Comparing YOLOv8 and Mask R-CNN for instance segmentation in complex orchard environments. Artificial Intelligence in Agriculture, 13, 84–99. https://doi.org/10.1016/j.aiia.2024.04.001 11. Liu, C., Zhang, R., & Guo, L. (2019). Part-pose guided Amur tiger re-identification. In 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW). https://doi.org/10.1109/ICCVW.2019.00042 12. Liu, N., Zhao, Q., Zhang, N., Cheng, X., & Zhu, J. (2019). Pose-guided complementary features learning for Amur tiger re-identification. In 2019 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). https://doi.org/10.1109/ICCVW.2019.00038
Статью можно приобрести в электронном виде (PDF формат).
Стоимость статьи 700 руб. (в том числе НДС 20%). После оформления заказа, в течение нескольких дней, на указанный вами e-mail придут счет и квитанция для оплаты в банке.
После поступления денег на счет издательства, вам будет выслан электронный вариант статьи.
Для заказа скопируйте doi статьи:
10.14489/vkit.2026.02.pp.013-022
и заполните форму
Отправляя форму вы даете согласие на обработку персональных данных.
.
This article is available in electronic format (PDF).
The cost of a single article is 700 rubles. (including VAT 20%). After you place an order within a few days, you will receive following documents to your specified e-mail: account on payment and receipt to pay in the bank.
After depositing your payment on our bank account we send you file of the article by e-mail.
To order articles please copy the article doi:
10.14489/vkit.2026.02.pp.013-022
and fill out the form
.
|