Convolutional neural networks trained with a developmental sequence of blurry to clear images reveal core differences between face and object processing

Jang, Hojin; Tong, Frank

doi:10.1167/jov.21.12.6

Cited by 14 publications

(29 citation statements)

References 71 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Reinforcing this claim, our findings are consistent with other recent studies indicating that the trajectory of human visual development has positive functional implications [8,31,32]. This body of research illustrates how a reduction in input complexity in early development [21,48], as well as the overall trajectory of development [3], can help facilitate learning within a variety of domains.…”

Section: Discussionsupporting

confidence: 91%

“…Indeed, contrary to our results, Avbers ˇek et al [30] reported that, for categorization of unblurred images, models trained on intact images performed better than models trained on blurred images. Similarly, Jang and Tong [32] did not observe an overall benefit in object categorization for models trained on blurred images as compared to models trained on unblurred images.…”

Section: Relation To Previous Workmentioning

confidence: 87%

“…As discussed earlier, several recent studies similar to ours have investigated how training with blurred images impacts learning in neural network models [30][31][32][33]. However, the central task we explored-basic-level object categorization-was not observed to benefit from low-pass image training in any of these reports.…”

Section: Relation To Previous Workmentioning

confidence: 87%

“…Vogelsang et al [31] found that experience with blurred images during early training of a CNN enhanced the ability to learn invariant spatial representations, which are central to adult-like configural face processing. Similarly, Jang and Tong [32] found that a network trained with initially blurred faces and objects was invariant across blur for face recognition, but not object recognition. This dissociation appears consistent with the claim that face recognition relies on somewhat different mechanisms from generic object recognition.…”

Section: Past Studies Using Cnns and Image Blurmentioning

confidence: 89%

See 3 more Smart Citations

Early experience with low-pass filtered images facilitates visual category learning in a neural network model

2023

View full text Add to dashboard Cite

Humans are born with very low contrast sensitivity, meaning that inputs to the infant visual system are both blurry and low contrast. Is this solely a byproduct of maturational processes or is there a functional advantage for beginning life with poor visual acuity? We addressed the impact of poor vision during early learning by exploring whether reduced visual acuity facilitated the acquisition of basic-level categories in a convolutional neural network model (CNN), as well as whether any such benefit transferred to subordinate-level category learning. Using the ecoset dataset to simulate basic-level category learning, we manipulated model training curricula along three dimensions: presence of blurred inputs early in training, rate of blur reduction over time, and grayscale versus color inputs. First, a training regime where blur was initially high and was gradually reduced over time—as in human development—improved basic-level categorization performance in a CNN relative to a regime in which non-blurred inputs were used throughout training. Second, when basic-level models were fine-tuned on a task including both basic-level and subordinate-level categories (using the ImageNet dataset), models initially trained with blurred inputs showed a greater performance benefit as compared to models trained exclusively on non-blurred inputs, suggesting that the benefit of blurring generalized from basic-level to subordinate-level categorization. Third, analogous to the low sensitivity to color that infants experience during the first 4–6 months of development, these advantages were observed only when grayscale images were used as inputs. We conclude that poor visual acuity in human newborns may confer functional advantages, including, as demonstrated here, more rapid and accurate acquisition of visual object categories at multiple levels.

show abstract

Section: Discussionsupporting

confidence: 91%

Section: Relation To Previous Workmentioning

confidence: 87%

Section: Relation To Previous Workmentioning

confidence: 87%

Section: Past Studies Using Cnns and Image Blurmentioning

confidence: 89%

See 2 more Smart Citations

Early experience with low-pass filtered images facilitates visual category learning in a neural network model

2023

View full text Add to dashboard Cite

show abstract

“…Moreover, certain visual experiences, such as low visual acuity at birth, may be particularly important for highlighting global shape information ( Cassia et al, 2002 ; Ostrovsky et al, 2009 ). Indeed, ANNs trained with a blurry-to-clear visual regimen showed improved performance for categories where global information is important, such as faces ( Jang and Tong, 2021 ; Vogelsang et al, 2018 ).…”

Section: Resultsmentioning

confidence: 99%

Perception of an object’s global shape is best described by a model of skeletal structure in human infants

Ayzenberg

Lourenco

2022

eLife

View full text Add to dashboard Cite

Categorization of everyday objects requires that humans form representations of shape that are tolerant to variations among exemplars. Yet, how such invariant shape representations develop remains poorly understood. By comparing human infants (6–12 months; N=82) to computational models of vision using comparable procedures, we shed light on the origins and mechanisms underlying object perception. Following habituation to a never-before-seen object, infants classified other novel objects across variations in their component parts. Comparisons to several computational models of vision, including models of high-level and low-level vision, revealed that infants’ performance was best described by a model of shape based on the skeletal structure. Interestingly, infants outperformed a range of artificial neural network models, selected for their massive object experience and biological plausibility, under the same conditions. Altogether, these findings suggest that robust representations of shape can be formed with little language or object experience by relying on the perceptually invariant skeletal structure.

show abstract

Butterfly effects in perceptual development: A review of the ‘adaptive initial degradation’ hypothesis

Vogelsang,

Pipa

et al. 2024

Developmental Review

View full text Add to dashboard Cite

Convolutional neural networks trained with a developmental sequence of blurry to clear images reveal core differences between face and object processing

Cited by 14 publications

References 71 publications

Early experience with low-pass filtered images facilitates visual category learning in a neural network model

Early experience with low-pass filtered images facilitates visual category learning in a neural network model

Perception of an object’s global shape is best described by a model of skeletal structure in human infants

Butterfly effects in perceptual development: A review of the ‘adaptive initial degradation’ hypothesis

Contact Info

Product

Resources

About