CREATING A FACIAL RECOGNITION SYSTEM BASED ON A NEURAL NETWORK THAT WORKS IN REAL TIME

Measuring Equipment and Metrology, Год журнала: 2025, Номер 86, С. 80 - 85

Опубликована: Янв. 1, 2025

The article is devoted to the theoretical foundations and practical aspects of facial recognition using neural networks. paper analyzes various approaches feature extraction from images, as well methods for training networks recognition. Special attention paid problems related variations in lighting, expressions, other factors affecting recog- nition accuracy. describes creation optimization a network real-time human face Emphasis placed on modern deep learning architectures, such Convolutional Neural Networks (CNN), which provide high accuracy processing speed. Key stages process are analyzed: data preprocessing, model training, use hardware accelerators (e.g., GPU), integration algorithms into software real-world applications. given challenges lighting variability, clothing changes, privacy protection. proposes effective performance evaluates their impact Experimental results demonstrate speed, allowing development suitable security systems, smartphones, smart devices.

Язык: Английский

On-orbit Satellite Hierarchical Anomaly Detection Using Causal Structure Learning

Advances in Space Research, Год журнала: 2024, Номер unknown

Опубликована: Сен. 1, 2024

Язык: Английский