An Investigation of Global and Local Visual Complexity and Their Impact on Cognitive Load

This research investigates whether complexity in global visual features has a stronger impact on cognitive load. It utilises the effect of visual complexity in visual stimuli that causes humans to underestimate the duration of time; to gauge the level of cognitive load in the visual process. Global and local visual complexity can be indicated by the image jpeg file size of a range of different Potts segmentation granularity coefficient γ. By exposing participants to images with different complexity in global and local features, the level of cognitive load becomes the indicator of which aspects of visual complexity are taking effect. In addition, studying eye movement and the ‘region of interest’ provides supportive information for analysing the visual process.

The experiment demonstrates that complexity in global features does generally take precedence in our cognitive process. In the case of photographic images, a numerical threshold for global features can be determined by identifying which value of granularity coefficient γ has the most significant impact on participants’ cognitive load measures. This helps us to understand which level of local details is effective in visual processing without specific detail image searching tasks. Which provides supportive evidence for the top-down visual processing theory.

The results also show the subjective complexity rating reported by participants has a higher correlation with cognitive load level than the objective numerical values of complexity. Implicate that the nature of complexity is inseparable from subjective biases. Both findings lead to a confirmation that semantic content predominately influences our cognitive process. These conclusions underscore the need for typological study on image semantic content, which may become necessary for further research on the nature of visual complexity.