Detection and Localization of HTML Presentation Failures Using Computer Vision-Based Techniques

An attractive and visually appealing appearance is important for the success of a website. Presentation failures in a site's web pages can negatively impact end users' perception of the quality of the site and the services it delivers. Debugging such failures is challenging because testers must visually inspect large web pages and analyze complex interactions among the HTML elements of a page. In this paper we propose a novel automated approach for debugging web page user interfaces. Our approach uses computer vision techniques to detect failures and can then identify HTML elements that are likely to be responsible for the failure. We evaluated our approach on a set of real-world web applications and found that the approach was able to accurately and quickly identify faulty HTML elements.