Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Deep convolutional neural networks in the face of caricature


Real-world face recognition requires us to perceive the uniqueness of a face across variable images. Deep convolutional neural networks (DCNNs) accomplish this feat by generating robust face representations that can be analysed in a multidimensional ‘face space’. We examined the organization of viewpoint, illumination, gender and identity in this space. We found that DCNNs create a highly organized face similarity structure in which identities and images coexist. Natural image variation is organized hierarchically, with face identity nested under gender, and illumination and viewpoint nested under identity. To examine identity, we caricatured faces and found that identification accuracy increased with the strength of identity information in a face, and caricature representations ‘resembled’ their veridical counterparts—mimicking human perception. DCNNs therefore offer a theoretical framework for reconciling decades of behavioural and neural results that emphasized either the image or the face in representations, without understanding how a neural code could seamlessly accommodate both.

A preprint version of the article is available at ArXiv.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Examples of face images.
Fig. 2: Visualization of the top-level DCNN similarity space for all images.
Fig. 3: Visualization of top-level similarity space with identity strength variation.
Fig. 4: Caricature effects.
Fig. 5: Density curves of face image-pair cosine similarity scores.

Data availability

All data used for analysis are available via the Open Science Framework at

Code availability

All of the code used for plotting and analysis is available via the Open Science Framework at


  1. 1.

    Marr, D. Vision: A Computational Investigation into the Human Representation and Processing of Visual Information (MIT Press, 1982).

  2. 2.

    Brunelli, R. & Poggio, T. Face recognition: features versus templates. IEEE Trans. Pattern Anal. Mach. Intell.15, 1042–1052 (1993).

    Article  Google Scholar 

  3. 3.

    Riesenhuber, M. & Poggio, T. Hierarchical models of object recognition in cortex. Nat. Neurosci.2, 1019–1025 (1999).

    Article  Google Scholar 

  4. 4.

    Bülthoff, H. H. & Edelman, S. Psychophysical support for a two-dimensional view interpolation theory of object recognition. Proc. Natl Acad. Sci. USA89, 60–64 (1992).

    Article  Google Scholar 

  5. 5.

    Yuille, A. L. Deformable templates for face recognition. J. Cogn. Neurosci.3, 59–70 (1991).

    Article  Google Scholar 

  6. 6.

    Biederman, I. Recognition-by-components: a theory of human image understanding. Psychol. Rev.94, 115–147 (1987).

    Article  Google Scholar 

  7. 7.

    Poggio, T. & Edelman, S. A network that learns to recognize three-dimensional objects. Nature343, 263–266 (1990).

    Article  Google Scholar 

  8. 8.

    Turk, M. & Pentland, A. Eigenfaces for recognition. J. Cogn. Neurosci.3, 71–86 (1991).

    Article  Google Scholar 

  9. 9.

    Valentine, T. A unified account of the effects of distinctiveness, inversion and race in face recognition. Q. J. Exp. Psychol. A43, 161–204 (1991).

    Article  Google Scholar 

  10. 10.

    Troje, N. F. & Bülthoff, H. H. Face recognition under varying poses: the role of texture and shape. Vision Res.36, 1761–1772 (1996).

    Article  Google Scholar 

  11. 11.

    O’Toole, A. J., Abdi, H., Deffenbacher, K. A. & Valentin, D. Low-dimensional representation of faces in higher dimensions of the face space. J. Opt. Soc. Am. A10, 405–411 (1993).

    Article  Google Scholar 

  12. 12.

    O’Toole, A. J., Deffenbacher, K. A., Valentin, D. & Abdi, H. Structural aspects of face recognition and the other-race effect. Mem. Cognit.22, 208–224 (1994).

    Article  Google Scholar 

  13. 13.

    Nestor, A., Plaut, D. C. & Behrmann, M. Feature-based face representations and image reconstruction from behavioral and neural data. Proc. Natl Acad. Sci. USA113, 416–421 (2016).

    Article  Google Scholar 

  14. 14.

    Blanz, V. & Vetter, T. A morphable model for the synthesis of 3D faces. In Proceedings of the 26th Annual Conference on Computer Graphics and Interactive Techniques 187–194 (ACM Press/Addison-Wesley, 1999).

  15. 15.

    Benson, P. J. & Perrett, D. I. Perception and recognition of photographic quality facial caricatures: implications for the recognition of natural images. Eur. J. Cogn. Psychol.3, 105–135 (1991).

    Article  Google Scholar 

  16. 16.

    Benson, P. J. & Perrett, D. I. Visual processing of facial distinctiveness. Perception23, 75–93 (1994).

    Article  Google Scholar 

  17. 17.

    Byatt, G. & Rhodes, G. Recognition of own-race and other-race caricatures: implications for models of face recognition. Vision Res.38, 2455–2468 (1998).

    Article  Google Scholar 

  18. 18.

    Lee, K., Byatt, G. & Rhodes, G. Caricature effects, distinctiveness and identification: testing the face-space framework. Psychol. Sci.11, 379–385 (2000).

    Article  Google Scholar 

  19. 19.

    Rhodes, G., Byatt, G., Tremewan, T. & Kennedy, A. Facial distinctiveness and the power of caricatures. Perception26, 207–223 (1997).

    Article  Google Scholar 

  20. 20.

    Rhodes, G., Brennan, S. & Carey, S. Identification and ratings of caricatures: implications for mental representations of faces. Cogn. Psychol.19, 473–497 (1987).

    Article  Google Scholar 

  21. 21.

    Lowe, D.G. Object recognition from local scale-invariant features. In Proceedings of the Seventh IEEE International Conference onComputer Vision Vol. 2, 1150–1157 (IEEE, 1999).

  22. 22.

    Dalal, N. & Triggs, B. Histograms of oriented gradients for human detection. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2005 Vol. 1, 886–893 (IEEE, 2005).

  23. 23.

    Ojala, T., Pietikainen, M. & Maenpaa, T. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns. IEEE Trans. Pattern Anal. Mach. Intell.24, 971–987 (2002).

    MATH  Article  Google Scholar 

  24. 24.

    Riesenhuber, M. & Poggio, T. Models of object recognition. Nat. Neurosci.3, 1199–1204 (2000).

    Article  Google Scholar 

  25. 25.

    Moghaddam, B., Jebara, T. & Pentland, A. Bayesian face recognition. Pattern Recognition33, 1771–1782 (2000).

    Article  Google Scholar 

  26. 26.

    Taigman, Y., Yang, M., Ranzato, M. & Wolf, L. Deepface: closing the gap to human-level performance in face verification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 1701–1708 (IEEE, 2014).

  27. 27.

    Sankaranarayanan, S., Alavi, A., Castillo, C. & Chellappa, R. Triplet probabilistic embedding for face verification and clustering. In Proceedings of the IEEE International Conference on Biometrics Theory, Applications and Systems 1–8 (IEEE, 2016).

  28. 28.

    Schroff, F., Kalenichenko, D. & Philbin, J. Facenet: a unified embedding for face recognition and clustering. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 815–823 (IEEE, 2015).

  29. 29.

    Chen, J.-C. et al. An end-to-end system for unconstrained face verification with deep convolutional neural networks. In Proceedings of the IEEE International Conference on Computer Vision Workshops 118–126 (IEEE, 2015).

  30. 30.

    Ranjan, R., Sankaranarayanan, S., Castillo, C. D. & Chellappa, R. An all-in-one convolutional neural network for face analysis. In 12th IEEE International Conference on Automatic Face and Gesture Recognition (FG 2017) 17–24 (IEEE, 2017).

  31. 31.

    Fukushima, K. Neocognitron: a hierarchical neural network capable of visual pattern recognition. Neural Netw.1, 119–130 (1988).

    Article  Google Scholar 

  32. 32.

    KrizhevskyA., SutskeverI. & HintonG. E. ImageNet classification with deep convolutional neural networks. Adv. Neural Inf. Proc. Sÿst.25, 1097–1105 (2012).

    Google Scholar 

  33. 33.

    Parde, C. J. et al. Face and image representation in deep CNN features. In 12th IEEE International Conference onAutomatic Face and Gesture Recognition (FG 2017) 673–680 (IEEE, 2017).

  34. 34.

    O’TooleA. J., CastilloC. D., PardeC. J., HillM. Q. & ChellappaR. Face space representations in deep convolutional neural networks. Trends Cogn. Sci.22, 794–809 (2018).

    Article  Google Scholar 

  35. 35.

    DiCarlo, J. J. & Cox, D. D. Untangling invariant object recognition. Trends Cogn. Sci.11, 333–341 (2007).

    Article  Google Scholar 

  36. 36.

    Hong, H., Yamins, D. L., Majaj, N. J. & DiCarlo, J. J. Explicit information for category-orthogonal object properties increases along the ventral stream. Nat. Neurosci.19, 613–622 (2016).

    Article  Google Scholar 

  37. 37.

    Yamins, D. L. & DiCarlo, J. J. Using goal-driven deep learning models to understand sensory cortex. Nat. Neurosci.19, 356–365 (2016).

    Article  Google Scholar 

  38. 38.

    Brennan, S. E. Caricature generator: the dynamic exaggeration of faces by computer. Leonardo18, 170–178 (1985).

    Article  Google Scholar 

  39. 39.

    Rhodes, G. Superportraits: Caricatures and Recognition (Psychology Press, 1997).

  40. 40.

    Leopold, D. A., O’Toole, A. J., Vetter, T. & Blanz, V. Prototype-referenced shape encoding revealed by high-level aftereffects. Nat. Neurosci.4, 89–94 (2001).

    Article  Google Scholar 

  41. 41.

    Van Der Maaten, L. Accelerating t-SNE using tree-based algorithms. J. Mach. Learn. Res.15, 3221–3245 (2014).

    MathSciNet  MATH  Google Scholar 

  42. 42.

    Grill-Spector, K. & Weiner, K. S. The functional architecture of the ventral temporal cortex and its role in categorization. Nat. Rev. Neurosci.15, 536–548 (2014).

    Article  Google Scholar 

  43. 43.

    Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M. & Boyes-Braem, P. Basic objects in natural categories. Cogn. Psychol.8, 382–439 (1976).

    Article  Google Scholar 

  44. 44.

    EberhardtS., CaderJ. G. & SerreT. How deep is the feature analysis underlying rapid visual categorization? Adv. Neural Inf. Proc. Syst.29, 1100–1108 (2016).

    Google Scholar 

  45. 45.

    Kietzmann, T. C. et al. The occipital face area is causally involved in facial viewpoint perception. J. Neurosci.35, 16398–16403 (2015).

    Article  Google Scholar 

  46. 46.

    Natu, V. S. et al. Dissociable neural patterns of facial identity across changes in viewpoint. J. Cogn. Neurosci.22, 1570–1582 (2010).

    Article  Google Scholar 

  47. 47.

    Grill-Spector, K. et al. Differential processing of objects under various viewing conditions in the human lateral occipital complex. Neuron24, 187–203 (1999).

    Article  Google Scholar 

  48. 48.

    Yue, X., Cassidy, B. S., Devaney, K. J., Holt, D. J. & Tootell, R. B. Lower-level stimulus features strongly influence responses in the fusiform face area. Cerebral Cortex21, 35–47 (2010).

    Article  Google Scholar 

  49. 49.

    Kay, K. N., Weiner, K. S. & Grill-Spector, K. Attention reduces spatial uncertainty in human ventral temporal cortex. Curr. Biol.25, 595–600 (2015).

    Article  Google Scholar 

  50. 50.

    Szegedy, C. et al. Intriguing properties of neural networks. Preprint at (2013).

  51. 51.

    Bansal, A., Castillo, C. D., Ranjan, R. & Chellappa, R. The do’s and don’ts for CNN-based face verification. In ICCV Workshops 2545–2554 (IEEE, 2017).

  52. 52.

    Ranjan, R. et al. A Fast and Accurate System for Face Detection, Identification, and Verification. In Proceedings of the IEEE Transactions on Biometrics, Behavior, and Identity Science 82–96 (IEEE, 2019)

  53. 53.

    Chen, J.-C., Patel, V. M. & Chellappa, R. Unconstrained face verification using deep CNN features. In IEEE Winter Conference on Applications of Computer Vision (WACV) 1–9 (IEEE, 2016).

  54. 54.

    Bansal, A., Nanduri, A., Castillo, C. D., Ranjan, R. & Chellappa, R. UMDFaces: an annotated face dataset for training deep networks. In IEEE International Joint Conference on Biometrics (IJCB) 464–473 (IEEE, 2017).

  55. 55.

    Guo, Y., Zhang, L., Hu, Y., He, X. & Gao, J. MS-Celeb-1M: a dataset and benchmark for large-scale face recognition. In European Conference on Computer Vision 87–102 (Springer, 2016).

  56. 56.

    He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 770–778 (IEEE, 2016).

  57. 57.

    van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Mach. Learn. Res.9, 2579–2605 (2008).

    MATH  Google Scholar 

  58. 58.

    Wattenberg, M., Viégas, F. & Johnson, I. How to use t-SNE effectively. Distill1, e2 (2016).

    Article  Google Scholar 

Download references


This work had funding support from the Intelligence Advanced Research Projects Activity (IARPA). This research is based on work supported by the Office of the Director of National Intelligence (ODNI) and IARPA (via R&D contract no. 2014-14071600012). The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the ODNI, IARPA or the US Government.

Author information




All authors were involved in the conceptualization and design of the methodology of the study. M.Q.H., C.D.C., R.R. and J.-C.C. handled software. The original draft of the manuscript was prepared by M.Q.H. and A.J.O. Review and editing were carried out by M.Q.H., C.J.P., Y.I.C., C.D.C., V.B. and A.J.O. Formal analysis, investigation and visualization were done by M.Q.H. and C.J.P., with validation by M.Q.H., Y.I.C., C.J.P. and C.D.C. Supervision and funding acquisition were handled by C.D.C. and A.J.O., with project administration by A.J.O.

Corresponding author

Correspondence to Matthew Q. Hill.

Ethics declarations

Competing interests

University of Maryland has filed a US patent application that covers portions of network A. R.R. and C.D.C. are co-inventors on this patent.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary figures, Supplementary tables, Supplementary methods and Supplementary discussion.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Hill, M.Q., Parde, C.J., Castillo, C.D. et al. Deep convolutional neural networks in the face of caricature. Nat Mach Intell 1, 522–529 (2019).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing