CURRENT PARADIGMS IN IMAGE RECOGNITION: THEORY, METRICS AND PRACTICAL APPLICATIONS
Keywords:
Pattern recognition, classification, quality metrics, neural networksAbstract
This article presents a systematic analysis of the evolution of image recognition methods – from classical approaches based on manually designed features to modern deep learning architectures. The theoretical foundations of representing objects as feature spaces are examined. The research results demonstrate the full life cycle of intelligent system development and justify the selection of specific methods based on operating conditions and available computational resources.
References
1. Бишоп, К. Распознавание образов и машинное обучение = Pattern Recognition and Machine Learning [Текст] / К. Бишоп. – Springer, 2006. – 738 с.
2. Гудфеллоу, Я. Глубокое обучение = Deep Learning [Текст] / Я. Гудфеллоу, И. Бенджио, А. Курвилль. – М. : ДМК Пресс, 2018 (ориг. изд. MIT Press). – 652 с.
3. Гонсалес, Р. Цифровая обработка изображений = Digital Image Processing [Текст] / Р. Гонсалес, Р. Вудс. – М. : Техносфера, 2012 (ориг. изд. Pearson). – 1104 с.
4. Хасти, Т. Элементы статистического обучения = The Elements of Statistical Learning [Текст] / Т. Хасти, Р. Тибширани, Дж. Фридман. – Springer, 2009. – 745 с.
5. Szeliski, R. Computer Vision: Algorithms and Applications [Text] / R. Szeliski. – Springer, 2010. – 812 p.
6. LeCun, Y. Deep learning [Text] / Y. LeCun, Y. Bengio, G. Hinton // Nature. – 2015. – Vol. 521. – P. 436–444.
7. Scikit-learn: Machine Learning in Python [Electronic resource] : official documentation. – URL: https://scikit-learn.org (date of access: 18.03.2026).
8. FastAPI framework [Electronic resource] : interactive API documentation. – URL: https://fastapi.tiangolo.com (date of access: 18.03.2026).
9. NumPy [Electronic resource] : fundamental scientific computing with Python. – URL: https://numpy.org (date of access: 18.03.2026).
10. Pillow (PIL Fork) [Electronic resource] : Python Imaging Library. – URL: https://python-pillow.org (date of access: 18.03.2026).
