Algorithms for Defining Visual Regions-of-Interest: Comparison with Eye Fixations

By Claudio M. Privitera, Lawrence W. Stark. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000

614 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Human-Computer Interaction

View Paper: https://www.semanticscholar.org/paper/60c2b03024f89c3f67d05d6b60d4ba1b032942c4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Human-Computer Interaction

Citations in the corpus

Automatic foveation for video compression using a neurobiological model of visual attention (2004, background): “, giving preference to skin color or facial features, under the assumption that human faces should always be present and given high priority [4], [22], or learning image processing algorithms that maximize overlap with human scanpaths on a specific set of images [23]) and, thus, are often not univer”
Clustering of Gaze During Dynamic Scene Viewing is Predicted by Motion (2011, methodology): “Privitera and Stark [100] compared actual and simulated foveations using k-means clustering.”
A principled approach to detecting surprising events in video (2005, methodology): “Three models are compared , illustrated in Figure 4: Shannon entropy as computed by Privitera and Stark [26] , saliency as computed by Itti and Koch [12], and surprise as presented here.”
Saliency, attention, and visual search: an information theoretic approach. (2009, neutral): “based definition of the sort described in Renninger et al. (2007) is that a definition based on minimizing local Journal of Vision (2009) 9(3):5, 1–24 Bruce & Tsotsos 6”
Components of bottom-up gaze allocation in natural images. (2005, background): “A number of studies have taken information-theoretical or statistical approaches to eye movements, asking how fixated regions differ from non-fixated regions (Krieger, Rentschler, Hauske, Schill, & Zetzsche, 2000; Privitera & Stark, 2000; Reinagel & Zador, 1999).”
Gradient-Based Structural Similarity for Image Quality Assessment (2006, methodology): “The gradient map is obtained by using the two Sobel masks and equation(3).”
A coherent computational approach to model bottom-up visual attention (2006, background): “The first one consists in comparing the first fixations of the scan paths whereas the second compares two fixation density maps [14], [34], [36], [ 37 ], [38], [39].”
Adopting Abstract Images for Semantic Scene Understanding (2016, background): “They allow for the direct study of how to infer high-level semantic information, since they remove the reliance on noisy low-level object, attribute and relation detectors, or the tedious hand-labeling of real images.”
A Benchmark of Computational Models of Saliency to Predict Human Fixations (2012, methodology): “Others have updated the Koch-Ullman model by adding features such as symmetry [59] or curvedness [80].”
State-of-the-Art in Visual Attention Modeling (2013, background): “The interested reader is referred to [22], [23], [127] for eye movement studies and [24]”
Real-time video abstraction (2006, background): “Based on this principle , several computational models of visual salience have been proposed [Privitera and Stark 2000; Itti and Koch 2001].”
Methods for comparing scanpaths and saliency maps: strengths and weaknesses (2012, methodology): “The Levenshtein distance is the most common way to compare scanpaths (Josephson & Holmes, 2002; Privitera & Stark, 2000) and has been widely used for assessing the usability of Web pages (Baccino, 2004).”