Dynamics of perceptual organization in complex visual search: The identification of self organized criticality with respect to visual grouping principles

Farkas Jozsef Attila, University of Southern Mississippi

Abstract

The current project applies modern quantitative theories of visual perception to examine the effect of the Gestalt Law of proximity on visual cognition. Gestalt Laws are spontaneous dynamic processes (Brunswik & Kamiya, 1953; Wertheimer, 1938) that underlie the principles of perceptual organization. These principles serve as mental short-cuts, heuristic rule-of-thumb strategies that shorten decision-making time and allow continuous, efficient processing and flow of information (Hertwig & Todd, 2002). The proximity heuristic refers to the observation that objects near each other in the visual field tend to be grouped together by the perceptual system (Smith-Gratto & Fisher, 1999). Proximity can be directly quantified as the distance between adjacent objects (inter-object distances) in a visual array. Recent studies on eye movements have revealed the interactive nature of self organizing dynamic processes in visual cognition (Aks, Zelinsky, & Sprott, 2002; Stephen, & Mirman, 2010). Research by Aks and colleagues (2002) recorded eye-movements during a complex visual search task in which participants searched for a target among distracters. Their key finding was that visual search patterns are not randomly distributed, and that a simple form of temporal memory exists across the sequence of eye movements. The objective of the present research was to identify how the law of proximity impacts visual search behavior as reflected in eye movement patterns. We discovered that 1) eye movements are fractal; 2) more fractality will result in decreased reaction time during visual search, and 3) fractality facilitates the improvement of reaction times over blocks of trials. Results were interpreted in view of theories of cognitive resource allocation and perceptual efficiency. The current research could inspire potential innovations in computer vision, user interface design and visual cognition.