Construction of GUI Elements Recognition Model for AI Testing based on Deep Learning

Nowadays, GUI testing ensures the quality of GUI software. GUI automated testing by recording coordinates and handles has the disadvantages of poor reusability and low portability, which reduces the quality of GUI software testing. Therefore, GUI testing needs a more efficient testing to meet the rapid development of GUI software. AI can make testing more efficient. But the precondition of GUI software testing is to identify the GUI elements. In this context, the present paper studies the intelligent recognition of GUI elements based on deep learning, develops uniform markup rules for GUI elements and creates datasets to train the GUI element recognition model. According to the selected object detection network, the GUI element recognition model is trained, and the results are analyzed and discussed. The results of this project not only provide technical support for GUI intelligent testing, but also provide research objects and data for intelligent system reliability.