User Interface (UI) testing has become a common practice for quality assurance in industrial mobile applications, and many automated tools are used for testing. However, existing research methods do not focus on effective feedback mechanisms to present testing outcomes. Therefore, this research proposes a DQCNN-based feedback mechanism for mobile application testing. Initially, computer screens are pre-processed using SAR-SRGAN and contour formation. Then, patterns are analysed using MFWKLST, followed by GUI element recognition using the YOLO approach. From the identified GUI elements, backgrounds are subtracted and elements are classified as text, image, and click action using the DB-CD-SCAN approach. Features are then extracted from the text and images, and important features are selected using CST. Meanwhile, coordinates are detected from click actions using CD, and robotic movement is assessed based on these coordinates. Finally, the selected features and robotic movements are provided to DQCNN to generate feedback, which is returned to the robotic movement. The proposed method achieved 99.07% accuracy.
Natarajan et al. (Wed,) studied this question.