The developmental trajectory of object recognition robustness: Children are like small adults but unlike big deep neural networks
Open accessAbstract
In laboratory object recognition tasks based on undistorted photographs, both adult humans and deep neural networks (DNNs) perform close to ceiling. Unlike adults', whose object recognition performance is robust against a wide range of image distortions, DNNs trained on standard ImageNet (1.3M images) perform poorly on distorted images. However, the last 2 years have seen impressive gains in DNN distortion robustness, predominantly achieved through ever-increasing large-scale datasets-orders of magnitude larger than ImageNet. Although this simple brute-force approach is very effective in achieving human-level robustness in DNNs, it raises the question of whether human robustness, too, is simply due to extensive experience with (distorted) visual input during childhood and beyond. Here we investigate this question by comparing the core object recognition performance of 146 children (aged 4-15 years) against adults and against DNNs. We find, first, that already 4- to 6-year-olds show remarkable robustness to image distortions and outperform DNNs trained on ImageNet. Second, we estimated the number of images children had been exposed to during their lifetime. Compared with various DNNs, children's high robustness requires relatively little data. Third, when recognizing objects, children-like adults but unlike DNNs-rely heavily on shape but not on texture cues. Together our results suggest that the remarkable robustness to distortions emerges early in the developmental trajectory of human object recognition and is unlikely the result of a mere accumulation of experience with distorted visual input. Even though current DNNs match human performance regarding robustness, they seem to rely on different and more data-hungry strategies to do so.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The formation of cortical representations of objects and concepts in the human brain spans across childhood into adolescence.
“In the human brain, the formation of cortical representations of objects and concepts spans across childhood into adolescence (e.g., see [81-83]).”
From How Adversarial REM Dreams May Facilitate Creativity, and Why We Become Aware of Them (2024) · cites this paper as [83]
The Introduction’s “Development of object recognition” section says the ventral visual cortex changes structurally and functionally from childhood through adolescence, and discusses cortical selectivity for object categories. It does not establish that cortical representations of concepts form across that period; indeed, it notes that children aged 5–12 already show adult-like object-category selectivity. Page numbers are unavailable in our copy. (LLM, conf=0.84)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 101 · Cited by 42
Research works printed but not linked: 18
- Bambach, S., Crandall, D., Smith, L., & Yu, C. (2018). Toddler-inspired visual object learning. 32nd Conference on Neural Information Processing Systems (NeurIPS).
- Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, accuracy: Quantifying trial-by-trial behaviour of CN worth 16x16 words: Transformers for image recognition at scale. 9th International Conference on Learning Representations (ICLR), 03–07 May 2021, Vienna, Austria.
- Gatys, L. A., Ecker, A. S., & Bethge, M. (2016). Image style transfer using convolutional neural networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2414– 2423.
- Geirhos, R., Meding, K., & Wichmann, F. A. (2020). Beyond accuracy: Quantifying trial-by-trial behaviour of CNNs and humans by measuring error consistency. 34th Conference on Neural Information Processing Systems (NeurIPS), 06–12 December, 2020, Vancouver, British Columbia, Canada.
- Geirhos, R., Narayanappa, K., Mitzkus, B., Thieringer, T., Bethge, M., Wichmann, F. A., et al. (2021). Partial success in closing the gap between human and machine vision. Advances in Neural Information Processing Systems, 34, 23885– 23899.
- Geirhos, R., Rubisch, P., Michaelis, C., Bethge, M., Wichmann, F. A., & Brendel, W. (2019). Imagenet-trained CNNs are biased towards texture; Increasing shape bias improves accuracy and robustness. International Conference on Learning Representations (ICLR), 06–09 May 2019, New Orleans, Louisiana, United States.
- Geirhos, R., Temme, C. R., Rauber, J., Schütt, H. H., Bethge, M., &Wichmann, F. A. (2018). Generalisation in humans and deep neural networks. 32nd Conference on Neural Information Processing Systems (NeurIPS), 31, 7538– 7550.
- Ha, D., & Schmidhuber, J. (2018). World models. arXiv preprint arXiv:1803.10122, 10.48550/arXiv.1803.10122. · DOI · Paper page
- Hafner, D., Lillicrap, T., Ba, J., & Norouzi, M. (2019). Dream to control: Learning behaviors by latent imagination. International Conference on Learning Representations (ICLR), 06–09 May 2019, New Orleans, Louisiana, United States.
- Karsh, R., & Breitenbach, F. W. (1983). Looking at looking: The amorphous fixation measure. In Eye movements and psychological functions (pp. 53– 64). New York: Routledge.
- Kolesnikov, A., Beyer, L., Zhai, X., Puigcerver, J., Yung, J., Gelly, S., et al. (2020). Big transfer (bit): General visual representation learning. European Conference on Computer Vision, 491– 507.
- Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25, 1097– 1105.
- Lin, L. J. (1991). Programming robots using reinforcement learning and teaching. AAAI, 781– 786.
- Mikołajczyk, A., & Grochowski, M. (2018). Data augmentation for improving deep learning in image classification problem. 2018 International Interdisciplinary PhD Workshop (IIPhDW), 09–12 May 2018, Swinoujscie, Poland, 117– 122.
- Orhan, E., Gupta, V., & Lake, B. M. (2020). Self-supervised learning through the eyes of a child. Advances in Neural Information Processing Systems, 33, 9960– 9971.
- Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. International Conference on Learning Representations (ICLR), April 14th to 16th 2014, Banff National Park, Canada.
- Xie, S., Girshick, R., Dollár, P., Tu, Z., & He, K. (2017). Aggregated residual transformations for deep neural networks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1492– 1500.
- Zhu, G., Zhang, M., Lee, H., & Zhang, C. (2020). Bridging imagination and reality for model-based deep reinforcement learning. Advances in Neural Information Processing Systems, 33, 8993– 9006.
- Authors
- Huber, L. S., Geirhos, R., & Wichmann, F. A.
- Affiliations
- University of Bern (CH)
- University of Tübingen (DE)
- ORCID (US)
- Corresponding author
- Huber, Lukas S.
- Geirhos, Robert
- Wichmann, Felix A.
- Venue
- Journal of Vision
- Published
- 2023-07-06
- Volume
- 23
- Issue
- 7
- Pages
- 4
- ISSN
- 1534-7362 (ISSN-L)
- PMC
- PMC10337805
- OpenAlex
- W4383301864
- Cited by
- 42
- References
- 101
- Citations by year
- 2026: 10 · 2025: 15 · 2024: 15 · 2023: 1 · 2022: 1
- FWCI
- 6.14
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- Journal of Vision (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Funders
- Deutsche Forschungsgemeinschaft (206/4-1, 390727645)
- Eberhard Karls Universität Tübingen (390727645)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
- Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- object recognition, image distortion robustness, deep neural network, human-machine comparison, shape cues, texture cues, visual experience, developmental trajectories
- MeSH terms
- Adult, Child, Humans, Visual Perception, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
