Analysis of Recognition Methods in National Flag Recognition Tasks
The paper presents a comparative analysis of the application of the leading object detection algorithms – YOLOv8 and Faster R-CNN, to the task of recognizing national flags. The subject of the study is the analysis of computer vision methods for accurate image recognition in real time. The purpose of the work is to conduct a comparative analysis of existing computer vision methods to determine the best indicators. The research methodology includes the use of a dataset of 14918 images for training and 1989 images for validation. To evaluate the effectiveness of algorithms, metrics such as accuracy, completeness, AP@0.5 , AP@0.5:0.95 and FPS. The results of the study showed that YOLOv8 provides high processing speed and good accuracy at low IoU, but has disadvantages in the form of reduced accuracy at high IoU and large model size. Faster R-CNN, on the other hand, provides high accuracy at high IoU and a smaller model size, but has disadvantages in the form of low processing speed and reduced accuracy at low IoU.
Сведения
- Журнал
- Trudy Universiteta
- Опубликовано
- 2024 · Т. 3, № 96
- Язык
- Английский
- ISSN
- 1609-1825 · 2710-3382
- Издатель
- Karaganda Technical University named after Abylkas Saginov