Scanpaths in Eye Movements during Pattern Perception

Noton, David; Stark, Lawrence

doi:10.1126/science.171.3968.308

Cited by 568 publications

(347 citation statements)

References 8 publications

(3 reference statements)

Supporting

Mentioning

322

Contrasting

Unclassified

Order By: Relevance

“…Hence, we do not expect perfect agreement between model-predicted salience and human eye position. In particular, our bottom-up model as used here does not yet account, among others, for how the rapid identification of the gist (semantic category) of a scene may provide contextual priors to more efficiently guide attention towards target objects of interest (Biederman, Teitelbaum, & Mezzanotte, 1983;Friedman, 1979;Hollingworth & Henderson, 1998;Oliva & Schyns, 1997;Potter & Levy, 1969;Torralba, 2003); how search for a specific target might be guided top-down, for example by boosting visual neurons tuned to the attributes of the target (Ito & Gilbert, 1999;Moran & Desimone, 1985;Motter, 1994;Müller, Reimann, & Krummenacher, 2003;Reynolds, Pasternak, & Desimone, 2000;Treue & Maunsell, 1996;Treue & Trujillo, 1999;Wolfe, 1994Wolfe, , 1998Wolfe, Cave, & Franzel, 1989;Yeshurun & Carrasco, 1998); or how task, expertise, and internal scene models may influence eye movements (Henderson & Hollingworth, 2003;Moreno, Reina, Luis, & Sabido, 2002;Nodine & Krupinski, 1998;Noton & Stark, 1971;Peebles & Cheng, 2003;Savelsbergh, Williams, van der Kamp, & Ward, 2002;Tanenhaus, Spivey-Knowlton, Eberhard, & Sedivy, 1995;Yarbus, 1967). Nevertheless, our hypothesis for this study is that a more realistic simulation framework might yield better agreement between human and model than a less realistic one.…”

mentioning

confidence: 99%

Quantitative modelling of perceptual salience at human eye position

Itti

2006

Visual Cognition

View full text Add to dashboard Cite

We investigate the extent to which a simple model of bottom-up attention and salience may be embedded within a broader computational framework, and compared with human eye movement data. We focus on quantifying whether increased simulation realism significantly affects quantitative measures of how well the model may predict where in video clips humans direct their gaze. We hence compare three variants of the model, tested with 15 video clips of natural scenes shown to three observers. We measure model-predicted salience at the locations gazed to by the observers, compared to random locations. The first variant simply processes the raw video clips. The second adds a gaze-contingent foveation filter. The third further attempts to realistically simulate dynamic human vision by embedding the video frames within a larger background, and shifting them to eye position. Our main finding is that increasing simulation realism significantly improves the predictive ability of the model. Better emulating the details of how a visual stimulus is captured by a constantly rotating retina during active vision has a significant positive impact onto quantitative comparisons between model and human behaviour.

show abstract

mentioning

confidence: 99%

Quantitative modelling of perceptual salience at human eye position

Itti

2006

Visual Cognition

View full text Add to dashboard Cite

show abstract

“…This measure enables the direct testing of hypotheses on the information search direction in risky decision making. Scanpath analysis, which focuses on the sequence of eye movements, provides spatiotemporal data on the spatial distribution of attention across a visual stimulus (Gbadamosi & Zangemeister, 2001;Noton & Stark, 1971;Underwood, Humphrey, & Foulsham, 2008;Zhou et al, 2013). The pattern generated by the scanpath enables researchers to investigate the process of risky choice using a global measure.…”

Section: Resultsmentioning

confidence: 99%

New Paradigms for the Old Question: Challenging the Expectation Rule Held by Risky Decision-Making Theories

Zhou

Zhang

et al. 2018

Journal of Pacific Rim Psychology

View full text Add to dashboard Cite

I n risky decision making, whether decision makers follow an expectation rule as hypothesised by mainstream theories is a compelling question. To tackle this question and enrich our knowledge of the underlying mechanism of risky decision making, we developed a series of new experimental paradigms that directly examined the computation processes to systematically investigate the process of risky decision making and explore the boundary condition of expectation rule over the course of a decade. In this article, we introduce these methods and review behavioural, eye-tracking, event-related potential, and functional magnetic resonance imaging studies that employed these methods. Results of these studies consistently showed that decision makers in the single-application condition did not perform the weighting and summing process assumed by the expectation rule. Moreover, decision makers were inclined to adopt a non-compensatory strategy, such as a heuristic one, in risky decision making. Furthermore, results indicated that the expectation rule was only applicable for conditions that involved decisions applied to numerous events (multiple applications) or to people (everyone). The findings indicated that using an index based on expected value to prescribe human risk preferences appears to be an artificial or false index of risk preference, and emphasised a new methodological direction for risky decision-making research.Keywords: risky decision making, process testing, compensatory rule, non-compensatory rule Decision making under risk is vital to human survival and development. Risky choices, such as investing or eating genetically modified foods, are common in everyday life. Determining how people make risky choices is a compelling question for scientists. For theoretical studies, a debate on whether decision making under risk is based on a compensatory or non-compensatory process is ongoing. Mainstream theories of decision making under risk claim that risky choices are based on a compensatory expectation-maximisation process. When making a risky choice, people weigh each possible outcome by its probability and then sum all the risky outcomes to assign an overall value (expectation) to each option. Then, they select the option that offers the highest overall expectation (Glöckner & Herbold, 2011;Kahneman & Tversky, 1979;Tversky & Kahneman, 1992). However, models that follow the non-compensatory rule and assume that people rely on only one (or a few) key dimension(s) rather than integrating information from all dimensions of an option Sciences, Beijing 100101, China. E-mail: lishu@psych.ac.cn; liangzy@psych.ac.cn to reach a decision are available (Brandstätter, Gigerenzer, & Hertwig, 2006; S. Li, 2004a;Thorngate, 1980). For example, the equate-to-differentiate model suggests that when making risky choices, people seek to 'equate' the less significant differences between options in either the best or worst possible payoff dimensions, and rely on only the large one-dimensional difference to be differentiated as the ...

show abstract

“…The list of vectors can now be searched for variances, which represent potentially interesting points (or regions). Indeed, our decomposition model can explain all variances discovered in seminal studies of human visual search (Noton, 1971;Treisman and Gormican, 1988). Others have already implemented architectures to mimic such search behavior (Itti et al, 1998;Privitera and Stark, 2000;Rajashekar et al, 2008), however those models merely extract straight contour orientations and can therefore explain only a small number of those findings.…”

Section: Further Comparison To Other Approachesmentioning

confidence: 99%

An Approach to the Parameterization of Structure for Fast Categorization

Rasche

2009

Int J Comput Vis

View full text Add to dashboard Cite

A decomposition is described, which parameterizes the geometry and appearance of contours and regions of gray-scale images with the goal of fast categorization. To express the contour geometry, a contour is transformed into a local/global space, from which parameters are derived classifying its global geometry (arc, inflexion or alternating) and describing its local aspects (degree of curvature, edginess, symmetry). Regions are parameterized based on their symmetric axes, which are evolved with a wave-propagation process enabling to generate the distance map for fragmented contour images. The methodology is evaluated on three image sets, the Caltech 101 set and two sets drawn from the Corel collection. The performance nearly reaches the one of other categorization systems for unsupervised learning.

show abstract

Scanpaths in Eye Movements during Pattern Perception

Cited by 568 publications

References 8 publications

Quantitative modelling of perceptual salience at human eye position

Quantitative modelling of perceptual salience at human eye position

New Paradigms for the Old Question: Challenging the Expectation Rule Held by Risky Decision-Making Theories

An Approach to the Parameterization of Structure for Fast Categorization

Contact Info

Product

Resources

About