In this chapter, the authors highlight the influence of visual saliency, or local contrast, on users’ searches of interfaces, particularly web pages. Designers have traditionally focused on the importance of goals and expectations (top-down processes) for the navigation of interfaces (Diaper & Stanton, 2004), with little consideration for the influence of saliency (bottom-up processes). The Handbook of Human-Computer Interaction (Sears & Jacko, 2008), for example, does not discuss the influence of bottom-up processing, potentially neglecting an important aspect of interface-based searches. The authors review studies that demonstrate how a user’s attention is rapidly drawn to visually salient locations in a variety of tasks and scenes, including web pages. They then describe an inexpensive, rapid technique that designers can use to identify visually salient locations in web pages, and discuss its advantages over similar methods.
TopIntroduction
Why is it that some interfaces guide us to the information we are searching for while others do not? In this chapter, we argue that visual saliency, a measure of an item’s visual uniqueness compared to its surroundings, plays a critical role in guiding our attention through an interface in cooperation with traditionally considered influences, such as goals. Visual saliency is inherent in natural and artificial scenes and interfaces, and cues the viewer to certain spatial regions over others. These cues may affect search difficulty. For instance, when a target is collocated with a salient region, searching can be accomplished in less time and with less effort and frustration for the user. The first objective of this chapter is to illustrate how visual saliency can be used to reduce user search times and aid task completion by implicitly guiding users to important information within an interface.
For an object to be visually salient, it must be visually unique relative to its surroundings (for a review see Healey, Booth, & Enns, 1996; Wolfe & Horowitz, 2004). For example, bolded and irregularity spaced text (s a l i e n t) amongst regularly-spaced, non-bolded text “pulls” the viewer’s attention making it more likely to be noticed. This aspect of salience is understood by the design community as illustrated by the use of bold, underlining, italics, color, and images to highlight important information. Making an item salient or identifying salient regions within an interface is unproblematic when the interface is primarily uniform. However, the same task is difficult in non-uniform interfaces, such as in web pages that are composed of pictures, logos, texts and varying spatial layouts. Thus, designers are often required to make a “best guess” about what is salient. The second objective of this chapter is to demonstrate that a saliency model, from the cognitive science literature, can help with the difficult but important task of determining which regions in a display are visually salient, and consequently the best locations for displaying critical information. In doing so, we will discuss research examining the processing of low-level cues that guide observers’ searches, as well as a computational model (Itti, Koch, & Niebur, 1998) that accurately predicts where users will look based purely on measures of visual saliency.
The saliency processes, which guide attention, can be demonstrated in relatively simple experimental displays. Further, these processes operate in essentially the same way regardless of one’s task: whether looking for a bird in the forest or a movie on a web page. Most of the research discussed in this chapter uses basic displays – natural scenes, experimental displays, web pages—and basic tasks—participants completing visual searches, or making responses to a target displayed among non-target (i.e., distractor) items.