Independence from Context Information Provided by Spatial Signature Learning in a Natural Object Localization Task

Giraudet, Guillaume; Roumes, Corinne

doi:10.1007/3-540-48315-2_14

Cited by 2 publications

(3 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…They find that at very short exposure times performance based on holistic information was superior, whereas the reverse was true when sufficient study time was available. Giraudet et al [22] use jumbled images to alter the contextual information in the image, and blurring to alter the image quality, to understand the cooperation between the top-down and bottom-up information. Top-down contextual information is important when the bottom-up appearance information is degraded.…”

Section: Localmentioning

confidence: 99%

Recognizing jumbled images: The role of local and global information in image classification

Parikh

2011

2011 International Conference on Computer Vision

View full text Add to dashboard Cite

The performance of current state-of-the-art computer vision algorithms at image classification falls significantly short as compared to human abilities. To reduce this gap, it is important for the community to know what problems to solve, and not just how to solve them. Towards this goal, via the use of jumbled images, we strip apart two widely investigated aspects: local and global information in images, and identify the performance bottleneck.Interestingly, humans have been shown to reliably recognize jumbled images. The goal of our paper is to determine a functional model that mimics how humans recognize jumbled images i.e. exploit local information alone, and further evaluate if existing implementations of this computational model suffice to match human performance. Surprisingly, in our series of human studies and machine experiments, we find that a simple bag-of-words based majorityvote-like strategy is an accurate functional model of how humans recognize jumbled images. Moreover, a straightforward machine implementation of this model achieves accuracies similar to human subjects at classifying jumbled images. This indicates that perhaps existing machine vision techniques already leverage local information from images effectively, and future research efforts should be focused on more advanced modeling of global information.

show abstract

Section: Localmentioning

confidence: 99%

Recognizing jumbled images: The role of local and global information in image classification

Parikh

2011

2011 International Conference on Computer Vision

View full text Add to dashboard Cite

show abstract

“…latéraux par exemple (Neisser, 1967 ;Lindsay et Norman, 1972). Ces deux catégories d'informations travaillent en collaboration de façon à ce que la décision soit prise dans les meilleurs délais et avec le maximum d'efficacité (Giraudet et Roumes, 1999Giraudet, 2000). Bon nombre de travaux se sont intéressés au problème de l'influence de la voie descendante ou cognitive, sur le traitement ascendant effectué par la voie sensorielle.…”

Section: Le Systè Me Visuel Et Les Images Naturellesunclassified

“…En effet, après une période d'entraînement et de familiarisation avec les stimulations floues, les sujets redeviennent capables de réaliser la tâche en utilisant uniquement les informations sensorielles disponibles, même si elles sont dégradées par rapport aux habitudes perceptives de l'observateur (Giraudet et Roumes, 1999Giraudet, 2000). (Giraudet et Roumes, 1999Giraudet, 2000 Habituellement, les informations de la voie sensorielle suffisent à détecter des objets cibles, les informations descendantes n'étant alors qu'optionnelles ou dévolues à des analyses plus complexes que la simple perception de l'objet (a) ; face à une contrainte, dans le cas présent un déficit des informations sensorielles, les système visuel réagit en accordant plus de poids aux données cognitives de contexte (b) ; puis, après une certaine période, il s'adapte à la contrainte et retourne à la situation initiale, malgré le flou (c) sont importantes. La plasticité cérébrale nécessaire à la mise en place de ce processus d'adaptation à des images dégradées est-elle présente chez des personnes plus âgées et, de surcroît, atteintes de DMLA ?…”

Section: Les Effets De L'apprentissageunclassified

Réapprendre à voir le monde : Rééduquer les personnes atteintes de dégénérescence maculaire en exploitant les capacités perceptives et cognitives de la vision périphérique

Giraudet

Corbe²,

Roumes

2005

Can. J. Aging

Self Cite

View full text Add to dashboard Cite

Age-related macular degeneration (ARMD) is a frequent cause of vision loss among people over age of 60. It is an aging process involving a progressive degradation of the central retina. It does not induce total blindness, since it does not affect the peripheral vision. Nonetheless, it makes difficult to read, drive, and perform all daily activities requiring fine details perception. Low-vision care consists in inducing an eccentric fixation so that relevant visual targets impact an unaffected retinal locus. It is necessary but not sufficient to enhance visual extraction. The present work aims to draw the attention of low-vision professionals to the necessity of developing new re-education tools. Beyond the perceptual re-education linked to an optimization of visual information extraction, a cognitive re-education should also be provided in order to enhance the interpretation processes. Indeed, the spatial-frequency properties of the visual world no longer match patient perceptual habits. The visually impaired person has to learn again to use these new sensory data in an optimal way. Contextual information can be a precious help in this learning process. An experimental study involving young people provides elements for another method of low-vision care, in terms of visual cognitive re-education.

show abstract

Independence from Context Information Provided by Spatial Signature Learning in a Natural Object Localization Task

Cited by 2 publications

References 16 publications

Recognizing jumbled images: The role of local and global information in image classification

Recognizing jumbled images: The role of local and global information in image classification

Réapprendre à voir le monde : Rééduquer les personnes atteintes de dégénérescence maculaire en exploitant les capacités perceptives et cognitives de la vision périphérique

Contact Info

Product

Resources

About