Логотип
ИЭА СТР
УДК 004.8 • EDN: WTZVUW • e-Library ID: 91674040

КОМПЛЕКСНАЯ ДИНАМИЧЕСКАЯ ОЦЕНКА КАЧЕСТВА СИСТЕМ КОМПЬЮТЕРНОГО ЗРЕНИЯ: МЕТОДИКА И ПРИНЦИПЫ ПОСТРОЕНИЯ

• С. 43-51
🇷🇺 На русском

Для цитирования

Шарова Д.Е. , Гарбук С.В. Комплексная динамическая оценка качества систем компьютерного зрения: методика и принципы построения // Информационноэкономические аспекты стандартизации и технического регулирования. 2026. № 3. С. 43–51.

Аннотация

Развитие систем с технологией искусственного интеллекта (ИИ), основанных на методах компьютерного зрения, сопровождается ростом требований к объективной, воспроизводимой и нормативно обоснованной оценке их качества. Традиционные подходы сосредоточены на показателях точности и не учитывают устойчивость алгоритмов, интерпретируемость, влияние на пользователя и динамику характеристик во времени. В результате оценка качества остается фрагментарной и не позволяет сформировать целостное представление о поведении системы в реальной эксплуатации. Целью исследования является разработка методики комплексной динамической оценки качества систем с ИИ, обеспечивающей сопоставимость, прослеживаемость и возможность анализа изменений во времени. В основу методики положено использование репрезентативного перечня существенных факторов эксплуатации, определяющих вариативность условий применения систем и позволяющих достоверно оценить их ключевые показатели качества (метрики) – точность, устойчивость к изменениям данных, прозрачность и доверие, влияние на пользователя и организационные процессы, а также экономическую эффективность и другие. Каждый показатель нормируется и взвешивается по значимости, после чего агрегируется в интегральный показатель качества Q, принимающий значения в диапазоне от 0 до 1. Предложенный подход обеспечивает возможность количественного анализа динамики качества и выявления деградации или улучшений на основе мониторинга показателя Q(t). Учет вариативности существенных факторов и достаточного объема тестовых данных позволяет оценивать репрезентативность испытаний и достоверность получаемых результатов. Результаты демонстрируют, что методика формирует основу для комплексного управления качеством систем ИИ, объединяя требования к продукту и процессам жизненного цикла. Она может применяться для сертификационных испытаний, постпроектного мониторинга и оценки эффективности эксплуатации интеллектуальных систем в различных прикладных областях.

Ключевые слова

искусственный интеллект компьютерное зрение оценка качества показатели качества существенные факторы эксплуатации

Об авторах

Шарова Д. Е.

Шарова Д. Е. — Начальник управления, Департамент информационных технологий города Москвы, ( Москва, Россия )

Гарбук С. В.

Гарбук С. В. — Кандидат технических наук, Директор, ФГБУН ВИНИТИ РАН, ( Москва, Россия )

Список литературы

  1. 1. Powers D.M.W. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation // Journal of Machine Learning Technologies. 2011. Vol. 2, no. 1. P. 37–63
  2. 2. Chicco D., Jurman G. The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation // BMC Genomics. 2020. Vol. 21, no. 1. P. 6.
  3. 3. Fawcett T. An introduction to ROC analysis // Pattern Recognition Letters. 2006. Vol. 27, no. 8. P. 861–874.
  4. 4. Garbuk S.V. Intellimetry as a Way to Ensure AI Trustworthiness // 2018 International Conference on Artificial Intelligence Applications and Innovations. IC-AIAI. 2018. P. 27–30.
  5. 5. Brodersen K.H., Ong C.S., Stephan K.E., Buhmann J.M. The balanced accuracy and its posterior distribution // In 2010 20th international conference on pattern recognition. IEEE. 2010. P. 3121–3124.
  6. 6. Chicco D., Tötsch N., Jurman G. The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation // BioData Mining. 2021. Vol. 14, no. 1. P. 13.
  7. 7. Warrens M. J. Cohen’s kappa is a weighted average // Statistical Methodology. 2010. Vol. 8, no. 6. P. 473–484.
  8. 8. Recht B., Roelofs R., Schmidt L., Shankar V. Do ImageNet Classifiers Generalize to ImageNet // Proceedings of the 36th International Conference on Machine Learning. 2019. Vol. 97. P. 5389–5400.
  9. 9. Taori R., Dave A., Shankar V. et al. Measuring Robustness to Natural Distribution Shifts in Image Classification // Advances in Neural Information Processing Systems. 2020. Vol. 33. P. 18583–18599.
  10. 10. Gulrajani I., Lopez-Paz D. In Search of Lost Domain Generalization. 2020. arXiv:2007.01434v1
  11. 11. Goodfellow I.J., Shlens J., Szegedy C. Explaining and Harnessing Adversarial Examples. 2014. arXiv:1412.6572v3
  12. 12. Lu J., Liu A., Dong F. et. al. Learning under Concept Drift: A Review // Transactions on Knowledge and Data Engineering. IEEE. 2019. Vol. 31, no. 12. P. 2346–2363.
  13. 13. Amershi S., Weld D., Vorvoreanu M. et al. Guidelines for Human-AI Interaction // CHI Conference on Human Factors in Computing Systems. 2019. No. 3. P. 1–13.
  14. 14. Sendak M., D’Arcy J., Kashyap S. et al. A Path for Translation of Machine Learning Products into Healthcare Delivery // EMJ Innovations. 2020.
  15. 15. Holstein K., Wortman Vaughan J., Daumé H. et al. Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? // CHI Conference on Human Factors in Computing Systems. 2019. No. 600. P. 1–16.
  16. 16. Buçinca Z., Malaya M., Glassman E. Proxy Tasks and Subjective Measures Can Be Misleading in Evaluating Explainable AI Systems. Proceedings of the AAAI Conference on Human Computation and Crowdsourcing. 2020. P. 454–464.
  17. 17. Lai V., Tan C. On Human Predictions with Explanations and Predictions without Explanations // Proceedings of the Conference on Fairness, Accountability, and Transparency. 2019. P. 29–38.
  18. 18. El Arab R.A., Al Moosa O.A. Systematic review of cost effectiveness and budget impact of artificial intelligence in healthcare // NPJ Digital Medicine. 2025. Vol. 8, no.1. P. 548.
  19. 19. Enholm I.M., Papagiannidis E., Mikalef P. et al. Artificial Intelligence and Business Value: a Literature Review // Information Systems Frontires. 2022. Vol. 24, pp. 1709–1734.
  20. 20. Ribeiro M., Singh S., Guestrin C. “Why Should I Trust You?” Explaining the Predictions of Any Classifier // KDD Conference Proceedings. 2016, pp. 1135–1144.
  21. 21. Lundberg S., Lee S. A Unified Approach to Interpreting Model Predictions // Proceedings of the 31st International Conference on Neural Information Processing Systems. 2017. P. 4768–4777.
  22. 22. Doshi-Velez F., Kim B. Towards A Rigorous Science of Interpretable Machine Learning. 2017. arXiv: 1702.08608.
  23. 23. Poursabzi-Sangdeh F., Goldstein D., Hofman J., et.al. Manipulating and Measuring Model Interpretability. CHI Conference on Human Factors in Computing Systems. 2018. Art. no. 237. P. 1–52.
  24. 24. Cabitza F, Campagner A, Balsano C. Bridging the «last mile» gap between AI implementation and operation: «data awareness» that matters. Annals of Translational Medicine. 2020. Vol. 8, no. 7. P. 501.
  25. 25. Webb G.I., Hyde R., Cao H. et.al. Characterizing concept drift. Data Mining and Knowledge Discovery. 2016. Vol. 30, no. 4. Р. 964–994.
  26. 26. De Lange M., Aljundi R., Masana M. et.al. A continual learning survey: Defying forgetting in classification tasks. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021. Vol. 44, no. 7. Р. 3366–3385.
  27. 27. Шарова Д.Е., Гарбук С.В. Методы оценки качества систем компьютерного зрения: эволюция подходов, тенденции и ограничения // Информационно-экономические аспекты стандартизации и технического регулирования. 2026. № 1(88). С. 35–41.
  28. 28. Vinayagasu B., Srivatsa S.K. Software Quality in Artificial Intelligence System. Journal of Information Technology. 2007. Vol. 6, no. 6. Р. 835–842.
  29. 29. Izurieta C., Reimanis D., O’Donoghue E. et al. A Generalized approach to the operationalization of Software Quality Models. PeerJ Computer Science. 2024. Vol. 10. P. e2357.
  30. 30. O’connor P.D., Kleyner A.V. Practical Reliability Engineering. John Wiley & Sons. 2012.
  31. 31. Kokol P. Software quality: A Historical and Synthetic Content Analysis. 2021. arXiv: 2106.14598.
  32. 32. Гарбук С.В. Качественный анализ рисков в жизненном цикле систем искусственного интеллекта / В кн. Безопасность России. Правовые, социально-экономические и научно-технические аспекты. Тематический блок «Национальная безопасность». Системная инженерия в проблемах национальной безопасности. Научн. рук. чл.-корр. РАН Н.А. Махутов. С. 643–659. – М.: МГОФ «Знание», 2025. – 898 с.
  33. 33. Липаев В.В. Сертификация программных средств. Учебник. – М.: СИНТЕГ, 2010. – 344 с.
🇬🇧 In English

METHODS FOR EVALUATING THE QUALITY OF COMPUTER VISION SYSTEMS: EVOLUTION OF APPROACHES, TRENDS, AND LIMITATIONS

For citation

Sharova D.E., Garbuk S.V. Methods for evaluating the auality of computer vision systems: methodology and design principles. Information and economic aspects of standardization and technical regulation. 2026; 43–51. (In Russ.)

Abstract

The development of artificial intelligence (AI) systems based on computer vision methods is accompanied by increasing demands for objective, reproducible, and regulation-compliant quality assessment. Traditional approaches focus primarily on accuracy metrics and do not account for algorithmic robustness, interpretability, user impact, or the temporal dynamics of system performance. As a result, quality assessment remains fragmented and does not provide a holistic understanding of system behavior under real operating conditions. The aim of this study is to develop a methodology for comprehensive dynamic quality assessment of AI-based systems that ensures comparability, traceability, and the ability to analyze changes over time. The methodology is based on the use of a representative set of significant operational factors that define the variability of system usage conditions and enable reliable evaluation of key quality indicators (metrics), including accuracy, robustness to data shifts, transparency and trustworthiness, user and organizational impact, and economic efficiency. Each indicator is normalized and weighted according to its significance, after which it is aggregated into an integral quality index Q, ranging from 0 to 1. The proposed approach enables quantitative analysis of quality dynamics and the detection of degradation or improvement through continuous monitoring of Q(t). Accounting for the variability of operational factors and ensuring sufficient test data volume allows for assessing the representativeness of testing and the reliability of the resulting evaluation. The results demonstrate that the methodology establishes a foundation for comprehensive quality management of AI systems by integrating both product-level and lifecycle-process requirements. It can be applied in certification testing, post-deployment monitoring, and evaluation of the operational effectiveness of intelligent systems across various application domains.

Keywords

artificial intelligence computer vision quality assessment quality indicators significant operational factors

About the authors

Sharova D. E.

Sharova D. E. — Head of Division, Department of Information Technologies of the City of Moscow, ( Moscow, Russia )

Garbuk S. V.

Garbuk S. V. — Ph.D. (Engineering), Director, All-Russian Institute of Scientific and Technical Information of the Russian Academy of Sciences (VINITI RAS), ( Moscow, Russia )

References

  1. 1. Powers D.M.W. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation. Journal of Machine Learning Technologies. 2011. Vol. 2, no. 1, pp. 37–63.
  2. 2. Chicco D., Jurman G. The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation. BMC Genomics. 2020. Vol. 21, no. 1. P. 6.
  3. 3. Fawcett T. An introduction to ROC analysis. Pattern Recognition Letters. 2006. Vol. 27, no. 8, pp. 861–874.
  4. 4. Garbuk S.V. Intellimetry as a Way to Ensure AI Trustworthiness. 2018 International Conference on Artificial Intelligence Applications and Innovations. ICAIAI. 2018., pp. 27–30.
  5. 5. Brodersen K.H., Ong C.S., Stephan K.E., Buhmann J.M. The balanced accuracy and its posterior distribution. In 2010 20th international conference on pattern recognition. IEEE. 2010, pp. 3121–3124.
  6. 6. Chicco D., Tötsch N., Jurman G. The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation. BioData Mining. 2021. Vol. 14, no. 1. P. 13.
  7. 7. Warrens M. J. Cohen’s kappa is a weighted average, Statistical Methodology. 2010. Vol. 8, no. 6, pp. 473–484.
  8. 8. Recht B., Roelofs R., Schmidt L., Shankar V. Do ImageNet Classifiers Generalize to ImageNet. Proceedings of the 36th International Conference on Machine Learning. 2019. Vol. 97, pp. 5389–5400.
  9. 9. Taori R., Dave A., Shankar V. et al. Measuring Robustness to Natural Distribution Shifts in Image Classification. Advances in Neural Information Processing Systems. 2020. Vol. 33, pp. 18583–18599.
  10. 10. Gulrajani I., Lopez-Paz D. In Search of Lost Domain Generalization. 2020. arXiv:2007.01434v1.
  11. 11. Goodfellow I.J., Shlens J., Szegedy C. Explaining and Harnessing Adversarial Examples. 2014. arXiv:1412.6572v3.
  12. 12. Lu J., Liu A., Dong F., et. al. Learning under Concept Drift: A Review. Transactions on Knowledge and Data Engineering. IEEE. 2019. Vol. 31, no. 12, pp. 2346–2363.
  13. 13. Amershi S., Weld D., Vorvoreanu M. et al. Guidelines for Human-AI Interaction. CHI Conference on Human Factors in Computing Systems. 2019. No. 3, pp. 1–13.
  14. 14. Sendak M., D’Arcy J., Kashyap S. et al. A Path for Translation of Machine Learning Products into Healthcare Delivery // EMJ Innovations. 2020.
  15. 15. Holstein K., Wortman Vaughan J., Daumé H., et al. Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? CHI Conference on Human Factors in Computing Systems. 2019. No. 600, pp. 1–16.
  16. 16. Buçinca Z., Malaya M., Glassman E. Proxy Tasks and Subjective Measures Can Be Misleading in Evaluating Explainable AI Systems // Proceedings of the AAAI Conference on Human Computation and Crowdsourcing. 2020, pp. 454–464.
  17. 17. Lai V., Tan C. On Human Predictions with Explanations and Predictions without Explanations // Proceedings of the Conference on Fairness, Accountability, and Transparency. 2019. P. 29–38.
  18. 18. El Arab R.A., Al Moosa O.A. Systematic review of cost effectiveness and budget impact of artificial intelligence in healthcare // NPJ Digital Medicine. 2025. Vol. 8, no.1. P. 548.
  19. 19. Enholm I.M., Papagiannidis E., Mikalef P. et al. Artificial Intelligence and Business Value: a Literature Review // Information Systems Frontires. 2022. Vol. 24, pp. 1709–1734.
  20. 20. Ribeiro M., Singh S., Guestrin C. “Why Should I Trust You?” Explaining the Predictions of Any Classifier // KDD Conference Proceedings. 2016, pp. 1135–1144.
  21. 21. Lundberg S., Lee S. A Unified Approach to Interpreting Model Predictions // Proceedings of the 31st International Conference on Neural Information Processing Systems. 2017, pp. 4768–4777.
  22. 22. Doshi-Velez F., Kim B. Towards A Rigorous Science of Interpretable Machine Learning. 2017. arXiv: 1702.08608.
  23. 23. Poursabzi-Sangdeh F., Goldstein D., Hofman J. et.al. Manipulating and Measuring Model Interpretability // CHI Conference on Human Factors in Computing Systems. 2018. Art. no. 237, pp. 1–52.
  24. 24. Cabitza F, Campagner A, Balsano C. Bridging the «last mile» gap between AI implementation and operation: «data awareness» that matters // Annals of Translational Medicine. 2020. Vol. 8, no. 7. P. 501.
  25. 25. Webb G.I., Hyde R., Cao H. et.al. Characterizing concept drift // Data Mining and Knowledge Discovery. 2016. Vol. 30, no. 4, pp. 964–994.
  26. 26. De Lange M., Aljundi R., Masana M. et.al. A continual learning survey: Defying forgetting in classification tasks // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021. Vol. 44, no. 7, рр. 3366–3385.
  27. 27. Sharova D.E., Garbuk S.V. Methods For Evaluating The Quality Of Computer Vision Systems: Evolution Of Approaches, Trends, And Limitations. Information and Economic Aspects of Standardization and Technical Regulation. 2026, no. 1(88), pp. 35–41. (In Russ.).
  28. 28. Vinayagasu B., Srivatsa S.K. Software Quality in Artificial Intelligence System // Journal of Information Technology. 2007. Vol. 6, no. 6, pp. 835–842.
  29. 29. Izurieta C., Reimanis D., O’Donoghue E. et al. A Generalized approach to the operationalization of Software Quality Models // PeerJ Computer Science. 2024. Vol. 10. P. e2357.
  30. 30. O’connor P.D.T., Kleyner A. Practical Reliability Engineering. Wiley. 2012.
  31. 31. Kokol P. Software quality: A Historical and Synthetic Content Analysis. arXiv: 2106.14598. 2021.
  32. 32. Garbuk S.V. Qualitative risk analysis in the life cycle of artificial intelligence systems. In: Security of Russia. Legal, socio-economic and scientific-technical aspects. Thematic section “National Security”. Systems engineering in national security problems. Scientific editor: N.A. Makhutov. Moscow: MGOF “Znanie”, 2025, pp. 643–659. 898 p. (in Russ.).
  33. 33. Lipaev V.V. Certification of software tools. Moscow: SINTEG Publ., 2010. 344 p. (In Russ.)