Computer Vision Technologies for Image Object Classification
Ochiq kirish: yoʻq
Annotatsiya
this paper presents modern approaches to object analysis and image classification using deep learning architectures VGG19, U-Net, and MobileNet. The features of each model are examined: high accuracy of feature extraction in VGG19, efficiency of image segmentation in U-Net, as well as compactness and computational resource optimization in MobileNet. A comparative analysis of model performance on open datasets was conducted, including Accuracy, Precision, Recall, and F1-score metrics. The results showed that the combined application of these architectures allows achieving a balance between classification accuracy, processing speed, and computational efficiency. The study demonstrates the prospects of using hybrid solutions for tasks in medicine, agriculture, video surveillance systems, and intelligent environmental monitoring.
Hali tarjima qilinmagan