Scholarlink

This paper evaluates and compares the effectiveness of YOLOv8 and Faster R-CNN models in detecting web User interface (UI) elements, focusing on their performance in dynamic and interactive environments. A custom UI with interactive filters-such as text input, dropdowns, sliders, and date pickers-was designed using React.js, and data was collected via Selenium for model training. The results reveal a clear speed-accuracy trade-off: YOLOv8 processes UI elements at 42 FPS, whereas Faster R-CNN achieves a higher mean Average Precision (mAP) of 0.91, particularly excelling in detecting complex elements like dropdowns (91%) and date pickers (85%). The study highlights the trade-offs between speed and precision, proposing a hybrid approach to leverage the strengths of both models for enhanced UI detection in automated testing frameworks.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Automated Testing: A Performance Analysis of YOLOv8 and Faster R-CNN for Dynamic Web UI Detection

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Safa Salem,
  • Nada Salaheddin Gheriyani,
  • Sohaib Osman,
  • Ayad Keshlaf,
  • Juma Ibrahim

摘要

This paper evaluates and compares the effectiveness of YOLOv8 and Faster R-CNN models in detecting web User interface (UI) elements, focusing on their performance in dynamic and interactive environments. A custom UI with interactive filters-such as text input, dropdowns, sliders, and date pickers-was designed using React.js, and data was collected via Selenium for model training. The results reveal a clear speed-accuracy trade-off: YOLOv8 processes UI elements at 42 FPS, whereas Faster R-CNN achieves a higher mean Average Precision (mAP) of 0.91, particularly excelling in detecting complex elements like dropdowns (91%) and date pickers (85%). The study highlights the trade-offs between speed and precision, proposing a hybrid approach to leverage the strengths of both models for enhanced UI detection in automated testing frameworks.