The prominence of behavioural biases in eye guidance

Tatler, Benjamin W.; Vincent, Ben

doi:10.1080/13506280902764539

Cited by 160 publications

(208 citation statements)

References 69 publications

(56 reference statements)

Supporting

Mentioning

199

Contrasting

Order By: Relevance

“…We are not the first to point this out, and research has begun to describe systematic biases and their causes during scene viewing in laboratory conditions (Tatler & Vincent, 2009;Foulsham & Underwood, 2008;Clarke & Tatler, 2014;Tseng et al, 2008). Clarke and Tatler (2014) have recently argued that -26-choosing an accurate central baseline is particularly important when assessing the performance of different models for predicting fixations in scenes.…”

Section: Discussionmentioning

confidence: 98%

Are fixations in static natural scenes a useful predictor of attention in the real world?

Foulsham

Kingstone

2017

Canadian Journal of Experimental Psychology / Revue canadienne

View full text Add to dashboard Cite

Word count (main body): 7275Unmasked Manuscript / Manuscrit non anonyme Click here to download Unmasked Manuscript / Manuscrit non anonyme StaticWalkView_Revision2.0.docx -2- AbstractResearch investigating scene perception normally proceeds in laboratory experiments using static images. Much has been learned about how observers look at pictures of the real world and the attentional mechanisms underlying this behaviour.However, the use of static, isolated pictures as a proxy for studying everyday attention in real environments has led to the criticism that such experiments are artificial. We report a new study that tests the extent to which the real world can be reduced to simpler laboratory stimuli. We recorded the gaze of participants walking on a university campus with a mobile eye tracker, and then showed static frames from this walk to new participants, in either a random or sequential order. The aim was to compare the gaze of participants walking in the real environment with fixations on pictures of the same scene.The data show that picture order affects inter-observer fixation consistency and changes looking patterns. Critically, while fixations on the static images overlapped significantly with the actual real-world eye movements, they did so no more than a model that assumed a general bias to the centre. Remarkably, a model that simply takes into account where the eyes are normally positioned in the head-independent of what is actually in the scene-does far better than any other model. These data reveal that viewing patterns to static scenes are a relatively poor proxy for predicting real world eye movement behaviour, while raising intriguing possibilities for how to best measure attention in everyday life.

show abstract

Section: Discussionmentioning

confidence: 98%

Are fixations in static natural scenes a useful predictor of attention in the real world?

Foulsham

Kingstone

2017

Canadian Journal of Experimental Psychology / Revue canadienne

View full text Add to dashboard Cite

show abstract

“…If we are interested in understanding naturalistic eye movement behavior, free viewing static images is not the most representative condition (14)(15)(16)(17)(18). Understanding image-based fixation behavior is not only a question of "where?…”

Section: Discussionmentioning

confidence: 99%

“…Accounting for the entirety of human eye movement behavior in naturalistic settings will require incorporating information about the task, high-level scene properties, and mechanistic constraints on the eye movement system (12,(15)(16)(17)(20)(21)(22). Our gold standard contains the influence of high-level (but still purely imagedependent) factors to the extent that they are consistent across observers.…”

Section: Discussionmentioning

confidence: 99%

Information-theoretic model comparison unifies saliency metrics

Kümmerer

Wallis

Bethge

2015

Proc. Natl. Acad. Sci. U.S.A.

147

167

View full text Add to dashboard Cite

Learning the properties of an image associated with human gaze placement is important both for understanding how biological systems explore the environment and for computer vision applications. There is a large literature on quantitative eye movement models that seeks to predict fixations from images (sometimes termed "saliency" prediction). A major problem known to the field is that existing model comparison metrics give inconsistent results, causing confusion. We argue that the primary reason for these inconsistencies is because different metrics and models use different definitions of what a "saliency map" entails. For example, some metrics expect a model to account for image-independent central fixation bias whereas others will penalize a model that does. Here we bring saliency evaluation into the domain of information by framing fixation prediction models probabilistically and calculating information gain. We jointly optimize the scale, the center bias, and spatial blurring of all models within this framework. Evaluating existing metrics on these rephrased models produces almost perfect agreement in model rankings across the metrics. Model performance is separated from center bias and spatial blurring, avoiding the confounding of these factors in model comparison. We additionally provide a method to show where and how models fail to capture information in the fixations on the pixel level. These methods are readily extended to spatiotemporal models of fixation scanpaths, and we provide a software package to facilitate their use.visual attention | eye movements | probabilistic modeling | likelihood | point processes H umans move their eyes about three times/s when exploring the environment, fixating areas of interest with the highresolution fovea. How do we determine where to fixate to learn about the scene in front of us? This question has been studied extensively from the perspective of "bottom-up" attentional guidance (1), often in a "free-viewing" task in which a human observer explores a static image for some seconds while his or her eye positions are recorded (Fig. 1A). Eye movement prediction is also applied in domains from advertising to efficient object recognition. In computer vision the problem of predicting fixations from images is often referred to as "saliency prediction," while to others "saliency" refers explicitly to some set of low-level image features (such as edges or contrast). In this paper we are concerned with predicting fixations from images, taking no position on whether the features that guide eye movements are "low" or "high" level.The field of eye movement prediction is quite mature: Beginning with the influential model of Itti et al. (1), there are now over 50 quantitative fixation prediction models, including around 10 models that seek to incorporate "top-down" effects (see refs. 2-4 for recent reviews and analyses of this extensive literature). Many of these models are designed to be biologically plausible whereas others aim purely at prediction (e.g., ref. 5). Progress is meas...

show abstract

“…Despite recent progress in understanding the factors that determine where an observer will eventually look in a scene, we know very little about what determines how an observer decides where he or she will specifically look next. Some evidence suggests that visual salience (Itti & Koch, 2000), occulomotor biases (Tatler & Vincent, 2009), the locations of prior fixations (Dodd, Van der Stigchel, & Hollingworth, 2009;Klein & MacInnes, 1999), semantic content (Underwood & Foulsham, 2006), and momentary task relevance (Hayhoe, Shrivastava, Mruczek, & Pelz, 2003) likely play roles, but identifying the factors that influence saccade sequencing remains an underdeveloped area of research that requires much greater scrutiny (Henderson, 2011). In this article, we consider the degree to which one's representation of a fixated object's spatio-temporal properties influences the direction in which a subsequent eye movement will travel.…”

mentioning

confidence: 99%

mentioning

confidence: 99%

Evaluating the influence of a fixated object’s spatio-temporal properties on gaze control

Cronin

Brockmole

2016

Atten Percept Psychophys

View full text Add to dashboard Cite

Despite recent progress in understanding the factors that determine where an observer will eventually look in a scene, we know very little about what determines how an observer decides where he or she will look next. We investigated the potential roles of object-level representations in the direction of subsequent shifts of gaze. In five experiments, we considered whether a fixated object's spatial orientation, implied motion, and perceived animacy affect gaze direction when shifting overt attention to another object. Eye movements directed away from a fixated object were biased in the direction it faced. This effect was not modified by implying a particular direction of inanimate or animate motion. Together, these results suggest that decisions regarding where one should look next are in part determined by the spatial, but not by the implied temporal, properties of the object at the current locus of fixation.Keywords Eye movements . Visual attention . Scene perception . Gaze direction . Spatial reference frames . Implied motion . AnimacyWe move our eyes three to four times each second as we investigate our visual surroundings. Each one of these eye movements requires a deceptively simple decision: Where should we look next? Despite recent progress in understanding the factors that determine where an observer will eventually look in a scene, we know very little about what determines how an observer decides where he or she will specifically look next. Some evidence suggests that visual salience (Itti & Koch, 2000), occulomotor biases (Tatler & Vincent, 2009) (Henderson, 2011). In this article, we consider the degree to which one's representation of a fixated object's spatio-temporal properties influences the direction in which a subsequent eye movement will travel.When viewing a scene, an observer forms several representations regarding objects' spatial position and temporal dynamics. For example, we use spatial reference frames to define objects' fronts, backs, left sides, right sides, and so forth. By doing so, we are able to describe a chair as Bfacing^a certain direction, being to the left of a table, or being behind a stool. In addition to defining the current spatial position of an object in a scene, it is also important to consider the ways in which an object's position might be altered over time. For example, some objects may be more or less likely to move. Similarly, considering the potential causes of an object's motion allows one to predict future events. Some objects may require an external force to act upon them, whereas others may be capable of volitional self-movement. One's assignment of reference frames (e.g., Carlson-Radvansky & Jiang, 1998;Logan, 1996;Logan & Sadler, 1996), representation of any possible motion (e.g., Gervais, Reed, Beall, & Roberts, 2010), and interpretation of animacy (e.g., Pratt, Radulescu, Guo, & Abrams, 2010) each affect covert aspects of attentional control. However, their impacts on overt attention (i.e., gaze control) have not been investigated.

show abstract

The prominence of behavioural biases in eye guidance

Cited by 160 publications

References 69 publications

Are fixations in static natural scenes a useful predictor of attention in the real world?

Are fixations in static natural scenes a useful predictor of attention in the real world?

Information-theoretic model comparison unifies saliency metrics

Evaluating the influence of a fixated object’s spatio-temporal properties on gaze control

Contact Info

Product

Resources

About