Adaptive Random Testing for Image Comparison in Regression Web Testing

Abstract

Web applications have become the most popular type of software in the past decade, attracting the attention of both the academia and the industry. In parallel with their popularity, the complexity of aesthetics and functionality of web applications have also increased significantly, creating a big challenge for maintenance and cross-browser compliance testing. Since such testing and verification activities require visual analysis, web application testing has not been sufficiently automated. In this paper, we propose a novel pairwise image comparison approach suitable for web application testing where the location of layout faults needs to be detected efficiently while insignificant variations being neglected. This technique is developed based on the characteristics of fault patterns of browser layouts. An empirical study conducted with the industry partner shows our approach is more effective and efficient than existing methods in this area.