Can (should) theories of crowding be unified?

Aga-Oglu, Mehmet; Chung, Susana T. L.

doi:10.1167/16.15.10

Cited by 23 publications

(20 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such a process is distinct from any single phenomenon such as source confusion, averaging or substitution but instead results from the broad spatial bandwidth of early stage filters. It is important to note that this model does not predict instances in which the magnitude of crowding is greatly modulated by certain configurations of flankers3361. We are not convinced, however, that these limitations argue against a population code; they instead require reconsideration of the way in which the features are weighted within the population code.…”

Section: Discussionmentioning

confidence: 89%

“…We previously asked participants to report both the target and flanker orientations in an experiment similar to the present report, and found that participants were generally capable of reporting both elements, though they often reversed feature positions41 (see also ref. 61). Taken together, these data reveal that perceptual reports in clutter are probabilistic, but relatively fine detail can be recovered.…”

Section: Discussionmentioning

confidence: 99%

“…Since submission of this article, Agaoglu and Chung61 published a study using a paradigm motivated by our earlier paper that investigates the validity of the present design. They levelled multiple criticisms of the continuous report approach and population response models of crowding.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Visual crowding is a combination of an increase of positional uncertainty, source confusion, and featural averaging

Harrison

Bex

2017

Sci Rep

View full text Add to dashboard Cite

Although we perceive a richly detailed visual world, our ability to identify individual objects is severely limited in clutter, particularly in peripheral vision. Models of such “crowding” have generally been driven by the phenomenological misidentifications of crowded targets: using stimuli that do not easily combine to form a unique symbol (e.g. letters or objects), observers typically confuse the source of objects and report either the target or a distractor, but when continuous features are used (e.g. orientated gratings or line positions) observers report a feature somewhere between the target and distractor. To reconcile these accounts, we develop a hybrid method of adjustment that allows detailed analysis of these multiple error categories. Observers reported the orientation of a target, under several distractor conditions, by adjusting an identical foveal target. We apply new modelling to quantify whether perceptual reports show evidence of positional uncertainty, source confusion, and featural averaging on a trial-by-trial basis. Our results show that observers make a large proportion of source-confusion errors. However, our study also reveals the distribution of perceptual reports that underlie performance in this crowding task more generally: aggregate errors cannot be neatly labelled because they are heterogeneous and their structure depends on target-distractor distance.

show abstract

Section: Discussionmentioning

confidence: 89%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Visual crowding is a combination of an increase of positional uncertainty, source confusion, and featural averaging

Harrison

Bex

2017

Sci Rep

View full text Add to dashboard Cite

show abstract

“…The parametric texture model of Portilla and Simoncelli (Portilla & Simoncelli, 2000;Simoncelli & Portilla, 1998) extended this approach by additionally matching the correlations between channels and other statistics, producing more realistic appearance matches to textures. This model has since had broad impact on the field of human perception and neuroscience: the texture statistic representation may provide a fruitful way to understand the processing in midventral visual areas (Freeman & Simoncelli, 2011;Freeman, Ziemba, Heeger, Simoncelli, & Movshon, 2013;Movshon & Simoncelli, 2014;Okazawa, Tajima, & Komatsu, 2015;Ziemba, Freeman, Movshon, & Simoncelli, 2016), and it has been argued to provide a good approximation of the type of information encoded in the periphery, and thus a model for tasks such as crowding and visual search (Balas, Nakano, & Rosenholtz, 2009;Freeman & Simoncelli, 2011;Keshvari & Rosenholtz, 2016;Rosenholtz, 2011;Rosenholtz, Huang, & Ehinger, 2012;Rosenholtz, Huang, Raj, Balas, & Ilie, 2012)-though other evidence questions the more general adequacy of this representation for explaining crowding and peripheral appearance (Agaoglu & Chung, 2016;Clarke, Herzog, & Francis, 2014;Herzog, Sayim, Chicherov, & Manassi, 2015;Wallis, Bethge, & Wichmann, 2016).…”

Section: Studying Texture Perception With Parametric Texture Modelsmentioning

confidence: 99%

A parametric texture model based on deep convolutional features closely matches texture appearance for humans

et al. 2017

View full text Add to dashboard Cite

Our visual environment is full of texture-"stuff" like cloth, bark, or gravel as distinct from "things" like dresses, trees, or paths-and humans are adept at perceiving subtle variations in material properties. To investigate image features important for texture perception, we psychophysically compare a recent parametric model of texture appearance (convolutional neural network [CNN] model) that uses the features encoded by a deep CNN (VGG-19) with two other models: the venerable Portilla and Simoncelli model and an extension of the CNN model in which the power spectrum is additionally matched. Observers discriminated model-generated textures from original natural textures in a spatial three-alternative oddity paradigm under two viewing conditions: when test patches were briefly presented to the near-periphery ("parafoveal") and when observers were able to make eye movements to all three patches ("inspection"). Under parafoveal viewing, observers were unable to discriminate 10 of 12 original images from CNN model images, and remarkably, the simpler Portilla and Simoncelli model performed slightly better than the CNN model (11 textures). Under foveal inspection, matching CNN features captured appearance substantially better than the Portilla and Simoncelli model (nine compared to four textures), and including the power spectrum improved appearance matching for two of the three remaining textures. None of the models we test here could produce indiscriminable images for one of the 12 textures under the inspection condition. While deep CNN (VGG-19) features can often be used to synthesize textures that humans cannot discriminate from natural textures, there is currently no uniformly best model for all textures and viewing conditions.

show abstract

“…The parametric texture model of Portilla and Simoncelli (Portilla & Simoncelli, 2000;Simoncelli & Portilla, 1998) extended this approach by additionally matching the correlations between channels and other statistics, producing more realistic appearance matches to textures. This model has since had broad impact on the field of human perception and neuroscience: the texture statistic representation may provide a fruitful way to understand the processing in mid-ventral visual areas (Freeman & Simoncelli, 2011;Freeman, Ziemba, Heeger, Simon-celli, & Movshon, 2013;Movshon & Simoncelli, 2014;Okazawa, Tajima, & Komatsu, 2015;Ziemba, Freeman, Movshon, & Simoncelli, 2016), and it has been argued to provide a good approximation of the type of information encoded in the periphery, and thus a model for tasks such as crowding and visual search (Balas, Nakano, & Rosenholtz, 2009;Freeman & Simoncelli, 2011;Keshvari & Rosenholtz, 2016;Rosenholtz, 2011;Rosenholtz, Huang, & Ehinger, 2012;Rosenholtz, Huang, Raj, Balas, & Ilie, 2012)-though other evidence questions the more general adequacy of this representation for explaining crowding and peripheral appearance (Agaoglu & Chung, 2016;Clarke, Herzog, & Francis, 2014;Herzog, Sayim, Chicherov, & Manassi, 2015;Wallis, Bethge, & Wichmann, 2016).…”

Section: Studying Texture Perception With Parametric Texture Modelsmentioning

confidence: 99%

A parametric texture model based on deep convolutional features closely matches texture appearance for humans

Wallis

Funke

Ecker

et al. 2017

Preprint

View full text Add to dashboard Cite

Our visual environment is full of texture-"stuff" like cloth, bark or gravel as distinct from "things" like dresses, trees or paths-and humans are adept at perceiving subtle variations in material properties. To investigate image features important for texture perception, we psychophysically compare a recent parameteric model of texture appearance (CNN model) that uses the features encoded by a deep convolutional neural network with two other models: the venerable Portilla and Simoncelli model (PS) and an extension of the CNN model in which the power spectrum is additionally matched. Observers discriminated model-generated textures from original natural textures in a spatial three-alternative oddity paradigm under two viewing conditions: when test patches were briefly presented to the near-periphery ("parafoveal") and when observers were able to make eye movements to all three patches ("inspection"). Under parafoveal viewing, observers were unable to discriminate 10 of 12 original images from CNN model images, and remarkably, the simpler PS model performed slightly better than the CNN model (11 textures). Under foveal inspection, matching CNN features captured appearance substantially better than the PS model (9 compared to 4 textures), and including the power spectrum improved appearance matching for two of the three remaining textures. None of the models we test here could produce indiscriminable images for one of the 12 textures under the inspection condition. While deep CNN (VGG-19) features can often be used to synthesise textures that humans cannot discriminate from natural textures, there is currently no uniformly best model for all textures and viewing conditions. Keywords: spatial vision, natural scenes, texture perception, peripheral vision . CC-BY 4.0 International license not peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was . http://dx.doi.org/10.1101/165761 doi: bioRxiv preprint first posted online IntroductionTextures are characterised by the repetition of smaller elements, sometimes with variation, to make up a pattern. Significant portions of the visual environment can be thought of as textures ("stuff" as distinct from "things"; Adelson & Bergen, 1991): your neighbour's pink floral wallpaper, the internal structure of dark German bread, the weave of a wicker basket, the gnarled bark of an old tree trunk, a bowl full of prawns ready for the barbie. Texture is an important material property whose perception is of adaptive value (Adelson, 2001;Fleming, 2014). For example, we can readily discriminate wet from dry stones (e.g. Ho, Landy, & Maloney, 2008), separating the underlying spatial texture from potentially temporary characteristics like glossiness. Where surfaces of different textures form occlusion boundaries, texture can provide a powerful segmentation cue; conversely, occlusion borders of similarly-textured surfaces can camoflage the occlusion (hiding a tiger among the leaves). Given the importance and ubiquity of visu...

show abstract

Can (should) theories of crowding be unified?

Cited by 23 publications

References 57 publications

Visual crowding is a combination of an increase of positional uncertainty, source confusion, and featural averaging

Visual crowding is a combination of an increase of positional uncertainty, source confusion, and featural averaging

A parametric texture model based on deep convolutional features closely matches texture appearance for humans

A parametric texture model based on deep convolutional features closely matches texture appearance for humans

Contact Info

Product

Resources

About