An image complexity measurement algorithm with visual memory capacity and an EEG study
Juan Huo · 2016 SAI Computing Conference (SAI) · 2016
A measurement method for the evaluation of the image complexity based on SIFT&K-means algorithm, namely the estimation of the mismatch between the target and the interesting points has been introduced in our previous research. Based on this method, we have made some improvements to calculate the image complexity of images with different memory targets. The improved algorithm SIFT&AIM&K-means was validated by the results from memory experiments and an EEG study. The image complexity value of this algorithm is not only consistent with human's visual sense but also can be applied to compare images with different size and memory contents. Besides the measurement method, the results also prove the hypothesis about the relationship between the image complexity, visual attention and visual working memory capacity again. This hypothesis indicates that the increased image complexity reduces the ability to focus on important visual information due to limited visual working memory capacity.