Automatically generating visually pleasing Information Walls

Abstract:

Information Wall (IW) is a GUI composed of rectangular cells each contains various kinds of information. The design and implementation of IW are mainly in a manual way nowadays, which typically leads to less variability in appearance style and low efficiency. This paper develops a tool to automate the generation of text-IWs based on a genetic programming algorithm. A model, which evaluates the visual quality of text-IWs, is also proposed and incorporated in the algorithm to make generated IWs more visually pleasing. Test results show that the tool is able to efficiently produce good looking IWs with variable appearance styles. With the help of the tool, designer productivity could be improved by utilizing machine-generated prototypes.