Description of Invention
This paper introduces an automated tool for detecting semantic elements in UML class diagrams using OpenCV image processing techniques. Designed for academic researchers specializing in UML diagrams, our tool identifies and extracts key features such as class counts, arrow types, and inheritance relationships. Unlike traditional machine learning methods, our approach employs rule-based image preprocessing, which does not require large training datasets and complex model training. Although our tool currently operates independently of machine learning, its powerful preprocessing capabilities lay a solid foundation for future integration with advanced learning algorithms. Such integration is expected to improve the accuracy and efficiency of UML class diagram analysis, showing the potential to significantly promote research and development in this field.