Design And Implementation Of Automatic Mobile Application Accessibility Detection System

Nowadays,Mobile Internet develops rapidly,the proportion of Internet users using mobile phones to access the Internet is as high as 99.1%.Mobile Internet has become the most important way for people to access information on the Internet.However,a large number of disabled people are unable to access network information because of the lack of support for accessibility in mobile applications.For example,the UI's component cannot be accessed by screen readers or the image contrast is too low,which will cause access barriers for the visually impaired.Small component will make it hard for the physically disabled to use.They will both bring challenges to the realization of information accessibility.To detect the accessibility support of mobile applications and enhance information accessibility in mobile apps,we design and implement an accessibility detection system for mobile application.The system obtains the app's page data of any third-party mobile apps through a reliable and general method,and trains the detection model based on machine learning to find out the accessibility problems of applications.Finally,the ability of the system to detect the accessibility problems is verified by experiments.The main contributions of this thesis are as follows:(1)This thesis proposes a method for automatic page data acquisition of mobile applications based on random traversal strategy.This method is general and can be applied to third-party applications to dynamically capture running app's final page information.(2)This thesis proposes a heuristic similarity comparison algorithm based on mobile application's structural information and image information.This algorithm can be used to remove application's similar pages to reduce data redundancy and improve system efficiency.(3)This thesis design and implement a mobile application accessibility detection system.Based on the machine learning method,our system trains the application's page structure detection model and the application's page image detection model to evaluate mobile application's accessibility support.We analyze the page structure and image of more than 15000 components on 8 popular mobile applications to verify the effectiveness of the above methods.The experimental results show that the average accuracy of the proposed system in detecting accessibility support of app's layout and image is up to 91.52% and 96.22% respectively.