A Computational Model to Predict Aesthetic Quality of Text Elements of GUI

The role of aesthetics in determining usability of interactive systems has come under focus in recent time. The issue is relevant for Graphical User Interfaces (GUI) containing elements of widely varying nature. It is important to evaluate GUI aesthetically to determine their acceptability to the users. Computational models have been reported in the literature to perform objective assessment of interface aesthetics. However, the existing models only consider geometric features at the highest level, without considering the content inside the geometry. To address this issue, we propose a computational model to evaluate aesthetics of textual contents present on a GUI. The proposed model is based on empirical data collected from user studies. The model is a weighted sum of six features characterizing text: chromatic contrast, luminance contrast, font size, letter spacing, line height and word spacing. A separate validation study demonstrates the feasibility and potential of the model (showing 87% accuracy in model prediction), which is expected to be useful in predicting usability of a web page in a more refined way. Such modeling has its obvious implications in the context of engineering interactive systems. The proposed model along with the user studies are presented in this paper.