Bootstrapped Evaluation with OctoDollop: A Mobile Application for Evaluating Mobile GUI Aesthetics in Context

There are currently many tools to evaluate the aesthetic quality of a graphical user interface, either by calculating metrics or by performing a machine learning analysis. Both methods typically work as follows: one takes a screenshot and then goes to a separate website or software to evaluate it, which induces a certain break between the context of use in which the screenshot was taken and the context of evaluation. Furthermore, metrics-based methods are often agnostic of the context of use while machine learning methods require a large size dataset to take the context into account. To facilitate the evaluation of a mobile graphical user interface being used in its real context of use, we demonstrate OctoDollop, a mobile application that instantly and seamlessly evaluates a graphical user interface without leaving its context of use and that is based on a limited number of samples by relying on an average bootstrapping. OctoDollop automatically segments the interface into regions, identifies its elements, and computes the overall aesthetic score towards harmony or contrast.