Survey of Methods Applying Deep Learning to Distinguish Between Computer Generated and Natural Images

  • Aiman MeenaiEmail author
  • Vasima Khan
Conference paper
Part of the Lecture Notes in Networks and Systems book series (LNNS, volume 100)


With the advent of fake news and propaganda being spread throughout the Internet using forged or computer-generated images, it is important to evolve algorithms that are able to differentiate between computer-generated images and natural ones. In this paper, we provide a high-level summary of the methods proposed recently which classify images as computer generated or natural using deep learning concepts. We spelled out the pros and cons of each method and further suggested future research paths like building a standard computer generated (CG) versus natural images (NI) dataset targeting compressed images from heterogeneous sources which ensures that the dataset models real world well, testing proposed approaches in real-life conditions and trying out training classifiers using combination of features generated from various convolutional neural networks (CNNs) as opposed to a single neural network to assess its impact on the accuracy rate for classification of compressed images.


Convolutional neural networks Image forensics Natural image Computer-generated image Deep learning Classification 


  1. 1.
    Broersma M, Graham T (2012) Social media as beat: tweets as a news source during the 2010 British and Dutch elections. Journalism Pract 6(3):403–419CrossRefGoogle Scholar
  2. 2.
    Hermida A (2010) Twittering the news: the emergence of ambient journalism. Journalism Pract 4(3):297–308CrossRefGoogle Scholar
  3. 3.
    Duggan M (2013) Photo and video sharing grow online. Pew Res Internet ProjGoogle Scholar
  4. 4.
    Aldiri K, Hobbs D, Qahwaji R (2008) The human face of e-business: engendering consumer initial trust through the use of images of sales personnel on e-commerce web sites. Int J E-Bus Res (IJEBR) 4(4):58–78CrossRefGoogle Scholar
  5. 5.
    Datta U, Sharma C (2013) Analysis of copy-move image forgery detection. Int J Adv Res Comput Sci Electron Eng (IJARCSEE) 2(8):607Google Scholar
  6. 6.
    Holmes O, Banks MS, Farid H (2016) Assessing and improving the identification of computer-generated portraits. ACM Trans Appl Percept (TAP) 13(2):7Google Scholar
  7. 7.
    Wu R, Li X, Yang B (2011) Identifying computer generated graphics via histogram features. In: 2011 18th IEEE international conference on image processing (ICIP). IEEEGoogle Scholar
  8. 8.
    Ng T-T, Chang S-F, Hsu J, Xie L, Tsui M-P (2005) Physics motivated features for distinguishing photographic images and computer graphics. In: Proceedings of ACM multimedia, pp 239–248Google Scholar
  9. 9.
    Dehnie S, Sencar T, Memon N (2006) Digital image forensics for identifying computer generated and digital camera images. In: Proceedings of IEEE ICIP, pp 2313–2316Google Scholar
  10. 10.
    Dirik AE, Bayram S, Sencar HT, Memon N (2007) New features to identify computer generated images. In: Proceedings of IEEE ICIP, vol IV, pp 433–436Google Scholar
  11. 11.
    Farid H, Lyu S (2003) Higher-order wavelet statistics and their application to digital forensics. In: CVPRW’03 conference on computer vision and pattern recognition workshop, 8. IEEEGoogle Scholar
  12. 12.
    Wang Y, Moulin P (2006) On discrimination between photorealistic and photographic images. Proc IEEE ICASSP II:161–164Google Scholar
  13. 13.
    Chen W, Shi YQ, Xuan G (2007) Identifying computer graphics using HSV color model and statistical moments of characteristic functions. In: Proceedings IEEEICME, pp 1123–1126Google Scholar
  14. 14.
    Sutthiwan P, Cai X, Shi YQ, Zhang H (2009) Computer graphics classification based on Markov process model and boosting feature selection technique. In: Proceedings IEEE ICIP, pp 2913–2916Google Scholar
  15. 15.
    Rahmouni N et al (2017) Distinguishing computer graphics from natural images using convolution neural networks. In: 2017 IEEE workshop on information forensics and security (WIFS). IEEEGoogle Scholar
  16. 16.
    Quan W et al (2018) Distinguishing between natural and computer-generated images using convolutional neural networks. IEEE Trans Inf Forensics Secur 13(11):2772–2787CrossRefGoogle Scholar
  17. 17.
    De Rezende ERS, Ruppert GCS, Carvalho T (2017) Detecting computer generated images with deep convolutional neural networks. In: 2017 30th SIBGRAPI conference on graphics, patterns and images (SIBGRAPI). IEEEGoogle Scholar
  18. 18.
    Srivastava N, Hinton GE, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15(1):1929–1958MathSciNetzbMATHGoogle Scholar
  19. 19.
    Ng T-T, Chang S-F, Hsu J, Pepeljugoski M (2004) Columbia photographicimagesandphotorealisticcomputergraphicsdataset. ADVENT, Columbia University, Tech. Rep. 205-2004-5Google Scholar
  20. 20.
    Yosinski J, Clune J, Bengio Y, Lipson H (2014) How transferable are features in deep neural networks? Adv Neural Inf Process Syst 3320–3328Google Scholar
  21. 21.
    Tokuda E, Pedrini H, Rocha A (2013) Computer generated images versus digital photographs: a synergetic feature and classifier combination approach. JVCI 24(8):1276–1292Google Scholar
  22. 22.
    Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. Adv Neural Inf Process Syst 1097–1105Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  1. 1.UIT-RGPVBhopalIndia
  2. 2.SIRTBhopalIndia

Personalised recommendations