Abstract: We proposed a  method for automatically identifying components with missing readable text in Android applications to improve the accessibility of these applications. Firstly, We used UI Automator to extract the graphical user interface of the application,  prune  irrelevant components and  complete  component attributes to generate the corresponding view tree. Secondly, we designed three heuristic rules  to identify components with missing readable text in the view tree. Through evaluation experiments on six popular applications, the proposed method successfully identified problematic components with an average accuracy  of 97%. Finally, a generated test report helped application developers  clearly locate and rectify the missing readable text by marking the problematic components in both the source code and screenshots. The  research achievement not only effectively improves the user experience for visually impaired users, enabling them to  interact with the applications more smoothly, but also provides developers with a practical tool to promote the overall accessibility of Android applications. Through  this approach, developers can better understand and address accessibility issues, creating a more user-friendly digital environment for all users.

Key words: accessibility, readable text, alternative text, problematic component recognition