Analysis of Image Complexity in Eyetracking Studies of Visual Art

Abstract

The article describes a method of analyzing the degree of complexity of images. Modification of images is often necessary when they differ with respect to statistical image properties such as complexity for instance. However, when it comes to studies on perception of art, such intervention might violate the actual structure of an artwork leading to severe methodological issues. We propose a different approach i.e. application of a coefficient aimed to control for unwanted effects of image complexity on eye movements. Present article describes the method of complexity coefficient calculation based on gradient counting for each pixel of an image. This approach might be useful in many studies that employ eyetracking to complex stimuli.