Academic journal article Perception and Psychophysics

The Role of First- and Second-Order Stimulus Features for Human Overt Attention

Academic journal article Perception and Psychophysics

The Role of First- and Second-Order Stimulus Features for Human Overt Attention

Article excerpt

When processing complex visual input, human observers sequentially allocate their attention to different subsets of the stimulus. What are the mechanisms and strategies that guide this selection process? We investigated the influence of various stimulus features on human overt attention-that is, attention related to shifts of gaze with natural color images and modified versions thereof. Our experimental modifications, systematic changes of hue across the entire image, influenced only the global appearance of the stimuli, leaving the local features under investigation unaffected. We demonstrated that these modifications consistently reduce the subjective interpretation of a stimulus as "natural" across observers. By analyzing fixations, we found that first-order features, such as luminance contrast, saturation, and color contrast along either of the cardinal axes, correlated to overt attention in the modified images. In contrast, no such correlation was found in unmodified outdoor images. Second-order luminance contrast ("texture contrast") correlated to overt attention in all conditions. However, although none of the second-order color contrasts were correlated to overt attention in unmodified images, one of the second-order color contrasts did exhibit a significant correlation in the modified images. These findings imply, on the one hand, that higher-order bottom-up effects-namely, those of second-order luminance contrast-may partially account for human overt attention. On the other hand, these results also demonstrate that global image properties, which correlate to the subjective impression of a scene being "natural," affect the guidance of human overt attention.

The complexity of natural input usually exceeds the parallel processing capacity of the human visual system. Consequently, the visual system sequentially attends to subsets of the input. Under natural viewing conditions, these shifts of attention are usually associated with changes in fixation. This so-called overt attention therefore provides an objective measure of attention shifts.

In the control of attention, two types of signals are to be distinguished, those related to the stimulus ("bottomup") and those related to subjects' expectations and experiences as well as to the task ("top-down"). Interest in the top-down features guiding human fixations dates at least back to Buswell's (1935) study. Comparing students of art with average observers, he found fixation durations to be slightly shorter for the "expert" group when looking at pictures; during reading (but not with pictures), a similar difference held true between adults and children. Both results highlight the roles of subjective experience and special training in human attention. In addition, Buswell provided some qualitative account of the effect of the task, which three decades later Yarbus's (1967) seminal study addressed in detail. Yarbus, in particular, demonstrated the dramatic influence of the task on overt attention in scenes that contain objects of high behavioral relevance to human observers, such as people or faces. On the basis of such results, there seems to be general consensus on the relative importance of experience and task to human attention. Nevertheless, bottom-up features are also assumed to play an important role for the allocation of spatial attention, especially in the absence of an explicit task. Consequently, the investigation of bottom-up models and their interaction with top-down signals provides fruitful insight into the mechanisms underlying human attention.

One of the most influential models to describe bottomup control of human attention is that of the so-called saliency map (Koch & Ullman, 1985). Though implementations of this model have undergone several modifications since its original formulation, the basic scheme remains unchanged (see Itti & Koch, 2001, for a review). The stimulus is analyzed in various feature channels, such as luminance, color, or orientation. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.