A measurement method for the mismatch between the image target and salient points as a metric for image complexity

In advertisement and web site design, a problem is the visual complexity caused mismatch between the target objects and the real salient objects. This mismatch can represent the degree of image complexity which is an important reason of low efficiency and unpleasant reading. Here this paper discusses this mismatch from different ways and introduces one new algorithm to measure the mismatch between the target objects and the real salient regions of an image. The algorithm combines the mathematic algorithm like SIFT(Scale Invariant Feature Transformation) and K-means with the cognitive science theory of visual working memory capacity. The measurement result of this algorithm can be a metric for the image complexity or visual complexity. The mismatch measured by this method has been validated by a visual experiment, which shows the SIFT&K-means algorithm is more approaching human´s visual sense compared to the other algorithm.