Automated Detection of TalkBack Interactive Accessibility Failures in Android Applications

TalkBack is one of the most popular screen readers for the 304 million visually impaired users worldwide to access Android devices. Yet, many popular Android apps lack a proper TalkBack accessible mechanism, which could cause failures that make apps unusable. Existing accessibility related techniques are limited in terms of detecting these issues. In this paper, we present a novel approach for automatically detecting TalkBack interactive accessibility failures that prevent TalkBack users from interacting with core functionalities in Android apps. Our evaluation of our technique on real-world Android apps showed that it is able to detect these TalkBack accessibility failures with high precision and recall.