<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">ntv</journal-id><journal-title-group><journal-title xml:lang="ru">Научно-технический вестник информационных технологий, механики и оптики</journal-title><trans-title-group xml:lang="en"><trans-title>Scientific and Technical Journal of Information Technologies, Mechanics and Optics</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2226-1494</issn><issn pub-type="epub">2500-0373</issn><publisher><publisher-name>Университет ИТМО</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.17586/2226-1494-2025-25-4-694-702</article-id><article-id custom-type="elpub" pub-id-type="custom">ntv-494</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>КОМПЬЮТЕРНЫЕ СИСТЕМЫ И ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>COMPUTER SCIENCE</subject></subj-group></article-categories><title-group><article-title>Исследование влияния состязательных атак на классификацию и кластеризацию изображений на примере модели ResNet50</article-title><trans-title-group xml:lang="en"><trans-title>The impact of adversarial attacks on a computer vision models perception of images Set intersection protocol with privacy preservation</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0004-5725-0526</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Болозовский</surname><given-names>Р. Р.</given-names></name><name name-style="western" xml:lang="en"><surname>Bolozovskii</surname><given-names>R. R.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Роман Ростиславович Болозовский, аспирант</p><p>197022; Санкт-Петербург</p></bio><bio xml:lang="en"><p>Roman R. Bolozovskii, PhD Student</p><p>197022; Saint Petersburg</p></bio><email xlink:type="simple">bolozovskii@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0003-4421-2411</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Левина</surname><given-names>А. Б.</given-names></name><name name-style="western" xml:lang="en"><surname>Levina</surname><given-names>A. B.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Алла Борисовна Левина, кандидат физико-математических наук, доцент, доцент кафедры</p><p>197022; Санкт-Петербург</p><p>sc 56427692900</p></bio><bio xml:lang="en"><p>Alla B. Levina, PhD (Physics &amp; Mathematics), Associate professor, Associate Professor at the Department</p><p>197022; Saint Petersburg</p><p>sc 56427692900</p></bio><email xlink:type="simple">Alla_levina@mail.ru</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0005-9232-5463</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Красов</surname><given-names>К. С.</given-names></name><name name-style="western" xml:lang="en"><surname>Krasov</surname><given-names>K. S.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Константин Сергеевич Красов, младший научный сотрудник</p><p>197022; Санкт-Петербург</p></bio><bio xml:lang="en"><p>Konstantin S. Krasov, Junior Researcher</p><p>197022; Saint Petersburg</p></bio><email xlink:type="simple">iflup@ya.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>Санкт-Петербургский государственный электротехнический университет «ЛЭТИ» им. В.И. Ульянова (Ленина)</institution><country>Россия</country></aff><aff xml:lang="en"><institution>Saint Petersburg Electrotechnical University “LETI”</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2025</year></pub-date><pub-date pub-type="epub"><day>29</day><month>08</month><year>2025</year></pub-date><volume>25</volume><issue>4</issue><fpage>694</fpage><lpage>702</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Болозовский Р.Р., Левина А.Б., Красов К.С., 2025</copyright-statement><copyright-year>2025</copyright-year><copyright-holder xml:lang="ru">Болозовский Р.Р., Левина А.Б., Красов К.С.</copyright-holder><copyright-holder xml:lang="en">Bolozovskii R.R., Levina A.B., Krasov K.S.</copyright-holder><license xml:lang="ru" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>Данная работа распространяется под лицензией Creative Commons Attribution 4.0.</license-p></license><license xml:lang="en" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://ntv.elpub.ru/jour/article/view/494">https://ntv.elpub.ru/jour/article/view/494</self-uri><abstract><sec><title>   Введение</title><p>   Введение. Прогресс в области компьютерного зрения привел к созданию мощных моделей, способных точно распознавать и интерпретировать визуальную информацию в различных областях знаний. На этом фоне растет уязвимость таких моделей к состязательным атакам — преднамеренному манипулированию входными данными с целью исказить модель машинного обучения и привести к неверным результатам распознавания. В работе приведены результаты исследования влияния различных типов состязательных атак на модель ResNet50 в задачах классификации и кластеризации изображений.</p></sec><sec><title>   Метод</title><p>   Метод. Исследованы следующие типы состязательных атак: метод быстрого градиентного знака, базовый итерационный метод, метод проецируемого градиентного спуска, метод Карлини и Вагнера, состязательная атака с использованием Elastic-Net, Expectation Over Transformation Predicted Gradient Descent, атаки на основе джиттера. Для визуализации областей внимания модели применен метод Gradient-Weighted Class Activation Mapping (Grad-CAM). Для визуализации кластеров в пространстве признаков использован алгоритм t-SNE. Устойчивость к атакам оценивалась по показателям успешности атак с использованием алгоритмов k-ближайших соседей иерархического маленького мира с различными метриками сходства.</p></sec><sec><title>   Основные результаты</title><p>   Основные результаты. Выявлены существенные различия в воздействии атак на внутренние представления модели и области фокусировки внимания. Показано, что итеративные методы атак вызывают значительные изменения в пространстве признаков и заметно влияют на визуализации Grad-CAM, тогда как простые атаки оказывают меньшее воздействие. Установлена высокая чувствительность большинства алгоритмов кластеризации к возмущениям. Наибольшую устойчивость среди исследованных подходов показала метрика внутреннего произведения.</p></sec><sec><title>   Обсуждение</title><p>   Обсуждение. Полученные результаты указывают на зависимость устойчивости модели от параметров атак и выбора метрик сходства, что проявляется в особенностях формирования кластерных структур. Выявленные закономерности в перераспределении пространства признаков в условиях целенаправленных атак открывают перспективы для дальнейшей оптимизации алгоритмов кластеризации, способных обеспечить более высокую степень защиты систем компьютерного зрения.</p></sec></abstract><trans-abstract xml:lang="en"><p>   Advances in computer vision have led to the development of powerful models capable of accurately recognizing and interpreting visual information in various fields of knowledge. However, these models are increasingly vulnerable to adversarial attacks – deliberate manipulations of input data designed to mislead the machine-learning model and produce incorrect recognition results. This article presents the results of an investigation into the impact of various types of adversarial attacks on the ResNet50 model in image classification and clustering tasks. Various types of adversarial attacks have been investigated: Fast Gradient Sign Method, Basic Iterative Method, Projected Gradient Descent, Carlini&amp;Wagner, Elastic-Net Attacks to Deep Neural Networks, Expectation Over Transformation Projected Gradient Descent, and jitter-based attacks. The Gradient-Weighted Class Activation Mapping (Grad-CAM) method was used to visualize the attention areas of the model. The t-SNE algorithm was applied to visualize clusters in the feature space. Attack robustness was assessed by attack success rate using k-Nearest Neighbors algorithm and Hierarchical Navigable Small World algorithms with different similarity metrics. Significant differences in the effects of attacks on the internal representations of the model and areas of focus have been identified. It is shown that iterative attack methods cause significant changes in the feature space and significantly affect Grad-CAM visualizations, whereas simple attacks have less impact. The high sensitivity of most clustering algorithms to perturbations has been established. The metric of the inner product showed the greatest stability among the studied approaches. The results obtained indicate the dependence of the stability of the model on the attack parameters and the choice of similarity metrics, which is manifested in the peculiarities of the formation of cluster structures. The observed feature-space redistributions under targeted attacks suggest avenues for further optimizing clustering algorithms to enhance the resilience of computer-vision systems.</p></trans-abstract><kwd-group xml:lang="ru"><kwd>состязательные атаки</kwd><kwd>компьютерное зрение</kwd><kwd>ResNet50</kwd><kwd>кластеризация изображений</kwd><kwd>KNN</kwd><kwd>HNSW</kwd></kwd-group><kwd-group xml:lang="en"><kwd>adversarial attacks</kwd><kwd>computer vision</kwd><kwd>ResNet50</kwd><kwd>image clustering</kwd><kwd>KNN</kwd><kwd>HNSW</kwd></kwd-group><funding-group><funding-statement xml:lang="ru">Работа выполнена в рамках государственного задания Министерства науки и высшего образования Российской Федерации № 075-00003-24-02 от 08. 02. 2024 (проект FSEE-2024-0003)</funding-statement><funding-statement xml:lang="en">The work was performed within the framework of the State Assignment of the Ministry of Science and Higher Education of the Russian Federation No. 075-00003-24-02 dated 08. 02. 2024 (FSEE-2024-0003 project)</funding-statement></funding-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Liu A. Guo J., Wang J., Liang S., Tao R., Zhou W., Liu C., Liu X., Tao D. X-adv: Physical adversarial object attacks against x-ray prohibited item detection // arXiv. 2023. arXiv:2302.09491. doi: 10.48550/arXiv.2302.09491</mixed-citation><mixed-citation xml:lang="en">Liu A. Guo J., Wang J., Liang S., Tao R., Zhou W., Liu C., Liu X., Tao D. X-adv: Physical adversarial object attacks against x-ray prohibited item detection. arXiv, 2023, arXiv:2302.09491. doi: 10.48550/arXiv.2302.09491</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Goodfellow I.J., Shlens J., Szegedy C. Explaining and harnessing adversarial examples // arXiv. 2015. arXiv:1412.6572. doi: 10.48550/arXiv.1412.6572</mixed-citation><mixed-citation xml:lang="en">Goodfellow I.J., Shlens J., Szegedy C. Explaining and harnessing adversarial examples. arXiv, 2015, arXiv:1412.6572. doi: 10.48550/arXiv.1412.6572</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Madry A., Makelov A., Schmidt L., Tsipras D., Vladu A. Towards deep learning models resistant to adversarial attacks // arXiv. 2019. arXiv:1706.06083. doi: 10.48550/arXiv.1706.06083</mixed-citation><mixed-citation xml:lang="en">Madry A., Makelov A., Schmidt L., Tsipras D., Vladu A. Towards deep learning models resistant to adversarial attacks. arXiv, 2019, arXiv:1706.06083. doi: 10.48550/arXiv.1706.06083</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">Carlini N., Wagner D. Towards evaluating the robustness of neural networks // Proc. of the IEEE Symposium on Security and Privacy (SP). 2017. P. 39–57. doi: 10.1109/SP.2017.49</mixed-citation><mixed-citation xml:lang="en">Carlini N., Wagner D. Towards evaluating the robustness of neural networks. Proc. of the IEEE Symposium on Security and Privacy (SP), 2017, pp. 39–57. doi: 10.1109/SP.2017.49</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Qian Y., He S., Zhao C., Sha J. Wang W., Wang B. Lea2: A lightweight ensemble adversarial attack via non-overlapping vulnerable frequency regions // Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV). 2023. P. 4487–4498. doi: 10.1109/iccv51070.2023.00416</mixed-citation><mixed-citation xml:lang="en">Qian Y., He S., Zhao C., Sha J. Wang W., Wang B. Lea2: A lightweight ensemble adversarial attack via non-overlapping vulnerable frequency regions. Proc. of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, pp. 4487–4498. doi: 10.1109/iccv51070.2023.00416</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Schlarmann C., Singh N.D., Croce F., Hein M. Robust CLIP: unsupervised adversarial fine-tuning of vision embeddings for robust large vision-language models // Proc. of the 41&lt;sup&gt;st&lt;/sup&gt; International Conference on Machine Learning. 2024. N 1779. P. 43684–43704.</mixed-citation><mixed-citation xml:lang="en">Schlarmann C., Singh N.D., Croce F., Hein M. Robust CLIP: unsupervised adversarial fine-tuning of vision embeddings for robust large vision-language models. Proc. of the 41&lt;sup&gt;st&lt;/sup&gt; International Conference on Machine Learning, 2024, no. 1779. pp. 43684–43704.</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">He K., Zhang X., Ren S., Sun J. Deep residual learning for image recognition // Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016. P. 770–778. doi: 10.1109/CVPR.2016.90</mixed-citation><mixed-citation xml:lang="en">He K., Zhang X., Ren S., Sun J. Deep residual learning for image recognition. Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778. doi: 10.1109/CVPR.2016.90</mixed-citation></citation-alternatives></ref><ref id="cit8"><label>8</label><citation-alternatives><mixed-citation xml:lang="ru">Liu X., Hu J., Yang Q., Jiang M., He J., Fang H. A divide-and-conquer reconstruction method for defending against adversarial example attacks // Visual Intelligence. 2024. V. 2. P. 30. doi: 10.1007/s44267-024-00061-y</mixed-citation><mixed-citation xml:lang="en">Liu X., Hu J., Yang Q., Jiang M., He J., Fang H. A divide-and-conquer reconstruction method for defending against adversarial example attacks. Visual Intelligence, 2024, vol. 2, pp. 30. doi: 10.1007/s44267-024-00061-y</mixed-citation></citation-alternatives></ref><ref id="cit9"><label>9</label><citation-alternatives><mixed-citation xml:lang="ru">Zhang J., Wu W., Huang J., Huang Y., Wang W., Su Y., Lyu M. Improving adversarial transferability via neuron attribution-based attacks // Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2022. P. 14973–14982. doi: 10.1109/CVPR52688.2022.01457</mixed-citation><mixed-citation xml:lang="en">Zhang J., Wu W., Huang J., Huang Y., Wang W., Su Y., Lyu M. Improving adversarial transferability via neuron attribution-based attacks. Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 14973–14982. doi: 10.1109/CVPR52688.2022.01457</mixed-citation></citation-alternatives></ref><ref id="cit10"><label>10</label><citation-alternatives><mixed-citation xml:lang="ru">Kurakin A., Goodfellow I., Bengio S. Adversarial examples in the physical world // Artificial Intelligence Safety and Security. 2018. P. 14. doi: 10.1201/9781351251389-8</mixed-citation><mixed-citation xml:lang="en">Kurakin A., Goodfellow I., Bengio S. Adversarial examples in the physical world. Artificial Intelligence Safety and Security, 2018, pp. 14. doi: 10.1201/9781351251389-8</mixed-citation></citation-alternatives></ref><ref id="cit11"><label>11</label><citation-alternatives><mixed-citation xml:lang="ru">Chen P.-Y., Sharma Y., Zhang H., Yi J. &amp; Hsieh C.-J. Ead: Elastic-net attacks to deep neural networks via adversarial examples // Proc. of the 32&lt;sup&gt;nd&lt;/sup&gt; AAAI Conference on Artificial Intelligence. 2018. V. 32. N 1. P. 10–17. doi: 10.1609/aaai.v32i1.11302</mixed-citation><mixed-citation xml:lang="en">Chen P.-Y., Sharma Y., Zhang H., Yi J. &amp; Hsieh C.-J. Ead: Elastic-net attacks to deep neural networks via adversarial examples. Proc. of the 32&lt;sup&gt;nd&lt;/sup&gt; AAAI Conference on Artificial Intelligence, 2018, vol. 32, no. 1, pp. 10–17. doi: 10.1609/aaai.v32i1.11302</mixed-citation></citation-alternatives></ref><ref id="cit12"><label>12</label><citation-alternatives><mixed-citation xml:lang="ru">Zimmermann R.S. Comment on “adv-bnn: Improved adversarial defense through robust bayesian neural network” // arXiv. 2019. arXiv:1907.00895. doi: 10.48550/arXiv.1907.00895</mixed-citation><mixed-citation xml:lang="en">Zimmermann R.S. Comment on “adv-bnn: Improved adversarial defense through robust bayesian neural network”. arXiv, 2019, arXiv:1907.00895. doi: 10.48550/arXiv.1907.00895</mixed-citation></citation-alternatives></ref><ref id="cit13"><label>13</label><citation-alternatives><mixed-citation xml:lang="ru">Schwinn L., Raab R., Nguyen A., Zanca D., Eskofier B. Exploring misclassifications of robust neural networks to enhance adversarial attacks // Applied Intelligence. 2023. V. 53. N 17. P. 19843–19859. doi: 10.1007/s10489-023-04532-5</mixed-citation><mixed-citation xml:lang="en">Schwinn L., Raab R., Nguyen A., Zanca D., Eskofier B. Exploring misclassifications of robust neural networks to enhance adversarial attacks. Applied Intelligence, 2023, vol. 53, no. 17, pp. 19843–19859. doi: 10.1007/s10489-023-04532-5</mixed-citation></citation-alternatives></ref><ref id="cit14"><label>14</label><citation-alternatives><mixed-citation xml:lang="ru">Selvaraju R.R., Cogswell M., Das A., Vedantam R., Parikh D., Batra D. Grad-cam: Visual explanations from deep networks via gradient-based localization // International Journal of Computer Vision. 2020. V. 128. N 2. P. 336–359. doi: 10.1007/s11263-019-01228-7</mixed-citation><mixed-citation xml:lang="en">Selvaraju R.R., Cogswell M., Das A., Vedantam R., Parikh D., Batra D. Grad-cam: Visual explanations from deep networks via gradient-based localization. International Journal of Computer Vision, 2020, vol. 128, no. 2, pp. 336–359. doi: 10.1007/s11263-019-01228-7</mixed-citation></citation-alternatives></ref><ref id="cit15"><label>15</label><citation-alternatives><mixed-citation xml:lang="ru">van der Maaten L., Hinton G. Visualizing data using t-SNE // Journal of Machine Learning Research. 2008. V. 9. P. 2579–2605.</mixed-citation><mixed-citation xml:lang="en">van der Maaten L., Hinton G. Visualizing data using t-SNE. Journal of Machine Learning Research, 2008, vol. 9, pp. 2579–2605.</mixed-citation></citation-alternatives></ref><ref id="cit16"><label>16</label><citation-alternatives><mixed-citation xml:lang="ru">Fix E., Hodges J. Discriminatory Analysis. Nonparametric Discrimination: Consistency Properties. USAF School of Aviation Medicine, 1951. 44 p.</mixed-citation><mixed-citation xml:lang="en">Fix E., Hodges J. Discriminatory Analysis. Nonparametric Discrimination: Consistency Properties. USAF School of Aviation Medicine, 1951, 44 p.</mixed-citation></citation-alternatives></ref><ref id="cit17"><label>17</label><citation-alternatives><mixed-citation xml:lang="ru">Malkov Y.A., Yashunin D.A. Efficient and robust approximate nearest neighbor search using hierarchical navigable small world graphs // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2020. V. 42. N 4. P. 824–836. doi: 10.1109/TPAMI.2018.2889473</mixed-citation><mixed-citation xml:lang="en">Malkov Y.A., Yashunin D.A. Efficient and robust approximate nearest neighbor search using hierarchical navigable small world graphs. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, vol. 42, no. 4, pp. 824–836. doi: 10.1109/TPAMI.2018.2889473</mixed-citation></citation-alternatives></ref><ref id="cit18"><label>18</label><citation-alternatives><mixed-citation xml:lang="ru">ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Stanford Vision Lab, Stanford University, Princeton University. ImageNet Data [Электронный ресурс]. URL: https://www.image-net.org/download.php. (дата обращения: 03. 03. 2025).</mixed-citation><mixed-citation xml:lang="en">ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Stanford Vision Lab, Stanford University, Princeton University. ImageNet Data [Электронный ресурс]. Available at: https://www.image-net.org/download.php. (accessed: 03. 03. 2025).</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
