Cargando…
Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations?
It has been suggested that perceiving blurry images in addition to sharp images contributes to the development of robust human visual processing. To computationally investigate the effect of exposure to blurry images, we trained convolutional neural networks (CNNs) on ImageNet object recognition wit...
Autores principales: | , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Frontiers Media S.A.
2023
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9975555/ https://www.ncbi.nlm.nih.gov/pubmed/36874839 http://dx.doi.org/10.3389/fpsyg.2023.1047694 |
_version_ | 1784898897178525696 |
---|---|
author | Yoshihara, Sou Fukiage, Taiki Nishida, Shin'ya |
author_facet | Yoshihara, Sou Fukiage, Taiki Nishida, Shin'ya |
author_sort | Yoshihara, Sou |
collection | PubMed |
description | It has been suggested that perceiving blurry images in addition to sharp images contributes to the development of robust human visual processing. To computationally investigate the effect of exposure to blurry images, we trained convolutional neural networks (CNNs) on ImageNet object recognition with a variety of combinations of sharp and blurred images. In agreement with recent reports, mixed training on blurred and sharp images (B+S training) brings CNNs closer to humans with respect to robust object recognition against a change in image blur. B+S training also slightly reduces the texture bias of CNNs in recognition of shape-texture cue conflict images, but the effect is not strong enough to achieve human-level shape bias. Other tests also suggest that B+S training cannot produce robust human-like object recognition based on global configuration features. Using representational similarity analysis and zero-shot transfer learning, we also show that B+S-Net does not facilitate blur-robust object recognition through separate specialized sub-networks, one network for sharp images and another for blurry images, but through a single network analyzing image features common across sharp and blurry images. However, blur training alone does not automatically create a mechanism like the human brain in which sub-band information is integrated into a common representation. Our analysis suggests that experience with blurred images may help the human brain recognize objects in blurred images, but that alone does not lead to robust, human-like object recognition. |
format | Online Article Text |
id | pubmed-9975555 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2023 |
publisher | Frontiers Media S.A. |
record_format | MEDLINE/PubMed |
spelling | pubmed-99755552023-03-02 Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? Yoshihara, Sou Fukiage, Taiki Nishida, Shin'ya Front Psychol Psychology It has been suggested that perceiving blurry images in addition to sharp images contributes to the development of robust human visual processing. To computationally investigate the effect of exposure to blurry images, we trained convolutional neural networks (CNNs) on ImageNet object recognition with a variety of combinations of sharp and blurred images. In agreement with recent reports, mixed training on blurred and sharp images (B+S training) brings CNNs closer to humans with respect to robust object recognition against a change in image blur. B+S training also slightly reduces the texture bias of CNNs in recognition of shape-texture cue conflict images, but the effect is not strong enough to achieve human-level shape bias. Other tests also suggest that B+S training cannot produce robust human-like object recognition based on global configuration features. Using representational similarity analysis and zero-shot transfer learning, we also show that B+S-Net does not facilitate blur-robust object recognition through separate specialized sub-networks, one network for sharp images and another for blurry images, but through a single network analyzing image features common across sharp and blurry images. However, blur training alone does not automatically create a mechanism like the human brain in which sub-band information is integrated into a common representation. Our analysis suggests that experience with blurred images may help the human brain recognize objects in blurred images, but that alone does not lead to robust, human-like object recognition. Frontiers Media S.A. 2023-02-15 /pmc/articles/PMC9975555/ /pubmed/36874839 http://dx.doi.org/10.3389/fpsyg.2023.1047694 Text en Copyright © 2023 Yoshihara, Fukiage and Nishida. https://creativecommons.org/licenses/by/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. |
spellingShingle | Psychology Yoshihara, Sou Fukiage, Taiki Nishida, Shin'ya Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title | Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title_full | Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title_fullStr | Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title_full_unstemmed | Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title_short | Does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
title_sort | does training with blurred images bring convolutional neural networks closer to humans with respect to robust object recognition and internal representations? |
topic | Psychology |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9975555/ https://www.ncbi.nlm.nih.gov/pubmed/36874839 http://dx.doi.org/10.3389/fpsyg.2023.1047694 |
work_keys_str_mv | AT yoshiharasou doestrainingwithblurredimagesbringconvolutionalneuralnetworksclosertohumanswithrespecttorobustobjectrecognitionandinternalrepresentations AT fukiagetaiki doestrainingwithblurredimagesbringconvolutionalneuralnetworksclosertohumanswithrespecttorobustobjectrecognitionandinternalrepresentations AT nishidashinya doestrainingwithblurredimagesbringconvolutionalneuralnetworksclosertohumanswithrespecttorobustobjectrecognitionandinternalrepresentations |