Performance Evaluation of Transfer Learning for Pornographic Detection

  • Buddhi Ashan
  • Hyuk Cho
  • Qingzhong LiuEmail author
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 1075)


The rapid growth of the Internet and online social network activities makes possible to fast share a huge volume of various digital multimedia contents such as images and videos. However, the easy accessibility of pornographic contents by general users, particularly teenagers, is problematic as well as the rare availability of the legitimate benchmark datasets for the pornographic image detection makes the research rather challenging. In this paper, we present a transfer learning approach, for which the existing general deep learning network is adopted for pornographic image detection problem. We consider five well-known deep learning models, which include VGG16, MobilNet, InceptionV3, Xception, and ResNet50, and each general deep learning network architecture is fine-turned to detect pornographic images. The experimental result with NPDI pornographic database demonstrates the effectiveness of the proposed transfer learning approach with promising detection accuracy.


Fine-tune Transfer learning VGG16 MobileNet InceptionV3 Xception ResNet Pornographic image detection Deep learning networks 



We would like to convey our gratitude to Professor Sandra Avila and her team for letting us use the NPDI dataset in a short notice.


  1. 1.
    Zhu, R., Wu, X., Zhu, B., Song, L.: Application of pornographic images recognition based on depth learning. In: Proceedings of the 2018 International Conference on Information Science and System - ICISS 2018 (2018).
  2. 2.
    Sengamedu, S., Sanyal, S., Satish, S.: Detection of pornographic content in internet images. In: Proceedings of the 19th ACM International Conference on Multimedia - MM 2011 (2011).
  3. 3.
    Zuo, H., Hu, W., Wu, O.: Patch-based skin color detection and its application to pornography image filtering. In: Proceedings of the 19th International Conference on World Wide Web - WWW 2010 (2010).
  4. 4.
    Platzer, C., Stuetz, M., Lindorfer, M.: Skin sheriff. In: Proceedings of the 2nd International Workshop on Security and Forensics in Communication Systems - SFCS 2014 (2014).
  5. 5.
    Karamizadeh, S., Arabsorkhi, A.: Methods of pornography detection. In: Proceedings of the 10th International Conference on Computer Modeling and Simulation - ICCMS 2018 (2018).
  6. 6.
    Jansohn, C., Ulges, A., Breuel, T.: Detecting pornographic video content by combining image features with motion information. In: Proceedings of the seventeen ACM international conference on Multimedia - MM 2009 (2009).
  7. 7.
    Avila, S., Thome, N., Cord, M., Valle, E., Araújo, A.D.: Pooling in image representation: the visual codeword point of view. Computer Vision and Image Understanding 117(5), 453–465 (2013). Scholar
  8. 8.
    Perez, M., et al.: Video pornography detection through deep learning techniques and motion information. Neurocomputing 230, 279–293 (2017). Scholar
  9. 9.
    Nian, F., Li, T., Wang, Y., Xu, M., Wu, J.: Pornographic image detection utilizing deep convolutional neural networks. Neurocomputing 210, 283–293 (2016). Scholar
  10. 10.
    Zhuo, L., Geng, Z., Zhang, J., Li, X.: ORB feature based web pornographic image recognition. Neurocomputing 173, 511–517 (2016). Scholar
  11. 11.
    Zhang, J., Sui, L., Zhuo, L., Li, Z., Yang, Y.: An approach of bag-of-words based on visual attention model for pornographic images recognition in compressed domain. Neurocomputing 110, 145–152 (2013)CrossRefGoogle Scholar
  12. 12.
    Yan, C., Liu, Y., Xie, H., Liao, Z., Yin, J.: Extracting salient region for pornographic image detection. J. Vis. Commun. Image Represent. 25(5), 1130–1135 (2014)CrossRefGoogle Scholar
  13. 13.
    Home - Keras Documentation (2019).
  14. 14.
    Simonyan, K., Zisserman, A.: Very Deep Convolutional Networks for Large-Scale Image Recognition, CoRR 14091556 (2014)Google Scholar
  15. 15.
    Howard, A., et al.: MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications, arXiv e-prints, arXiv:1704.04861 (2017)
  16. 16.
    Szegedy, C., et al.: Going Deeper with Convolutions, arXiv e-prints, arXiv:1409.4842 (2014)
  17. 17.
    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the Inception Architecture for Computer Vision, arXiv e-prints, arXiv:1512.00567 (2015). Accessed 18 Apr 2019
  18. 18.
    Francois, C.: Xception: Deep Learning with Depthwise Separable Convolutions, arXiv e-prints, arXiv:1610.02357 (2016)
  19. 19.
    Kaiming, H., Xiangyu, Z., Shaoqing, R., Jian, S.: Deep Residual Learning for Image Recognition, arXiv e-prints, arXiv:1512.03385 (2015)
  20. 20.
    Jin, X., Wang, Y., Tan, X.: Pornographic Image Recognition via Weighted Multiple Instance Learning arXiv e-prints, arXiv:1902.03771v1 (2019)CrossRefGoogle Scholar
  21. 21.
    Jung, J., Makhijani, R., Morlot, A.: Combining CNNs for detecting pornography in the absence of labeled training data (2017)Google Scholar
  22. 22.
    Ren, S., He, K., Girshick, R., Sun, J.: Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 39(6), 1137–1149 (2017). Scholar
  23. 23.
    Girshick, R.: Fast R-CNN. In: ICCV, pp. 1440–1448 (2015)Google Scholar
  24. 24.
    Huang, Y., Kong, A.W.K.: Using a CNN ensemble for detecting pornographic and upskirt images. In: 2016 IEEE 8th International Conference on Biometrics Theory, Applications and Systems (BTAS) (2016)Google Scholar
  25. 25.
    Lienhart, R., Hauke, R.: Filtering adult image content with topic models. In: IEEE International Conference on Multimedia and Expo. ICME 2009, pp. 1472–1475 (2009)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Department of Computer ScienceSam Houston State UniversityHuntsvilleUSA

Personalised recommendations