Article Preview
Top1. Introduction
Human visual attention is a cognitive process deployed to reduce the complexity of visual scene analysis (Wolfe, Visual Attention, 2000). For this purpose, a subset of the available visual information is selected by shifting the focus of attention across the visual scene to the most salient objects, locations, and features (Bisley, 2011). It is because of visual attention mechanisms that the human visual system is able to cope with the abundant amount of visual information that it is confronted with at any instant in time. Understanding visual attention and eye movement mechanisms (Engelke, Maeder, & Zepernick, 2009), relating them to human cognition (Wang, 2014) and applying them in cognitive computing applications has been recognised as a valuable tool to improve system performance (Itti, Koch, & Niebur, 1998) (Le Meur, Le Callet, Barba, & Thoreau, 2006) (Zhang, Tong, Marks, Shan, & Cottrell, 2008) (Engelke, Kaprykowsky, Zepernick, & Ndjiki-Nya, 2011).
Wang et al. define attention as “a perceptive process of the brain that focuses the mind or the perceptive engine on external stimuli, internal motivations, and/or threads of thought by selective concentration and proper responses.” (Wang, Patel, & Patel, 2013). We here look at selective attention, which is characterised by turning our focus towards a specific stimulus when there are two or more stimuli present in the current field of view (Harada, Mori, Yoshizawa, & Mizoguchi, 2014). Visual attention is driven by two main mechanisms:
- •
Overt attention relates to the process of actively directing a sense towards an item or event. In terms of overt visual attention, this is exhibited as moving the eye gaze towards an object or location in visual space. Saccadic eye movements facilitate fast shifts of the focus of attention across the scene and high spatial sampling enables to perceive the foveal regions with high accuracy.
- •
Covert attention describes the mechanism of mentally shifting the focus of attention without moving the eyes. For instance, it is possible to gaze at a certain object while attending objects in the visual periphery. Humans do this frequently, consciously as well as sub-consciously, especially in stimulus intensive scenarios such as driving a car through a busy street. For handling such visually complex scenarios we need to be able to quickly and reliably process visual information and attend to multiple stimuli in the visual field (Duenser & Mancero, 2009).
The relation between overt and covert attentional processes has been studied widely in the literature (Posner, 1980) (Hunt & Kingstone, 2003) (De Haan, Morgan, & Rorden, 2008). Most studies evolve around the paradigm that covert attention enhances stimulus appearance in peripheral vision. Posner et al. (Posner, 1980) have shown this in their earlier work through cueing experiments.
One area that heavily involves overt and covert attention is visual search (Wolfe, Cave, & Franzel, 1989) (Duncan & Humphreys, 1989) (McPeek & Keller, 2002). In its simplest definition, visual search denotes the task of finding a target amongst a set of distractors. This is typically done through an active scan of the environment by moving the eye gaze to potential target locations. Hence, eye movements have frequently been used to study visual search performance. Measuring eye movements, however, was found not to provide sufficiently deep insight into visual search performance as visual search is largely driven by pre-attentive (parallel) mechanisms in addition to serial search through eye movements.