Complex Conditional Generative Adversarial Nets for Multiple Objectives Detection in Aerial Images

  • Dan PopescuEmail author
  • Loretta Ichim
  • Andrei Docea
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11304)


Simultaneously detection and evaluation of small regions of interest from aerial images is successfully achieved by conditional generative adversarial nets (cGAN). As novelty, the paper proposes a cheap and accurate method based on a cGAN structure, containing two generators, and graphics processing units (GPU) to segment small flooded areas and respectively, roads from images taken by unmanned aerial vehicles. In the learning phase the weights for the discriminator and the two generators are established by a back propagation method. The real mask is created by using information of the color components R, G, B, H and a voting scheme in a supervised process. A set of 40 images were used for the learning phase and another set of 100 images were used for method validation. The method presents the advantages of accuracy and time processing (especially in the operational phase).


Aerial images Flood segmentation Road segmentation GPU image processing Conditional generative adversarial nets 



The work has been funded by project CAMIA, GEX 25/2017 (UPB), MAARS 185/2017 (ROSA) and MUWI 1224/2018 (NETIO).


  1. 1.
    Yulianto, F., Sofan, P., Zubaidah, A., Sukowati, K.A.D., Pasaribu, J.M., Khomarudin, M.R.: Detecting areas affected by flood using multi-temporal ALOS PALSAR remotely sensed data in Karawang, West Java. Indonesia. Nat. Hazards 77, 959–985 (2015)CrossRefGoogle Scholar
  2. 2.
    Pentari, A., Moirogiorgou, K., Livanos, G., Iliopoulou, D., Zervakis, M.: Feature analysis on river flow video data for floating tracers detection. In: Proceedings of the IEEE International Conference on Imaging Systems and Techniques (IST), 14–17 Octover, Santorini Island, Greece, pp. 287–292 (2014)Google Scholar
  3. 3.
    Lee, J.N., Kwak, K.C.: A trends analysis of image processing in unmanned aerial vehicle. Int. J. Comput. Inf. Sci. Eng. 8, 261–264 (2014)Google Scholar
  4. 4.
    Popescu, D., Ichim, L., Stoican, F.: Unmanned aerial vehicle systems for remote estimation of flooded areas based on complex image processing. Sensors 17(3), 446 (2017)CrossRefGoogle Scholar
  5. 5.
    Popescu, D., Ichim, L.: Aerial image segmentation by use of textural features. In: 20th International Conference on System Theory, Control and Computing (ICSTCC), 13–15 October, Sinaia, Romania, pp. 721–726 (2016)Google Scholar
  6. 6.
    Strigl, D., Kofler, K., Podlipnig, S.: Performance and scalability of GPU-based convolutional neural networks. In: 18th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP), 17–19 February, Pisa, Italy, pp. 1–8 (2010)Google Scholar
  7. 7.
    Goodfellow, I.J., et al.: Generative Adversarial Networks. arXiv:1406.2661v1 (2014)
  8. 8.
    Isola, P., Zhu, J.-Y., Zhou, T., Efros A.A.: Image-to-image translation with Conditional Adversarial Networks. arXiv:1611.07004v1 (2016)
  9. 9.
    Mirza, M., Osindero, S.: Conditional Generative Adversarial nets. arXiv:1411.1784v1 (2014)
  10. 10.
    Kataoka, Y., Matsubara, T., Uehara, K.: Image generation using Generative Adversarial Networks and attention mechanism. In: IEEE/ACIS 15th International Conference on Computer and Information Science (ICIS), 26–29 June, Okayama, Japan (2016)Google Scholar
  11. 11.
    Kohl, S., et al.: Adversarial Networks for the detection of aggressive prostate cancer. arXiv:1702.08014v1 (2017)
  12. 12.
    Wolterink, J.M., Leiner, T., Viergever, M.A., Isgum, I.: Generative Adversarial Networks for noise reduction in low-dose CT. IEEE Trans. Med. Imaging PP(99) (2017)Google Scholar
  13. 13.
    Tachibana, R., Matsubara,T., Uehara, K.: Semi-supervised learning using adversarial networks. In: IEEE/ACIS 15th International Conference on Computer and Information Science (ICIS), 26–29 June, Okayama, Japan (2016)Google Scholar
  14. 14.
    Yazdani, M.: RemixNet: generative adversarial networks for mixing multiple inputs. In: Semantic Computing (ICSC), 30 January–1 February (2017)Google Scholar
  15. 15.
    Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. In: 3rd International Conference for Learning Representations, San Diego, pp. 1–15 (2015)Google Scholar
  16. 16.
    Sumalan, A.L., Popescu, D., Ichim, L.: Flood evaluation in critical areas by UAV. In: 8th International Conference on Electronics, Computers and Artificial Intelligence (ECAI), 30 June–2 July, pp. 1–6 (2016)Google Scholar
  17. 17.
    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: MIC-CAI, pp. 234–241 (2015)Google Scholar
  18. 18.
    Torch | Scientific computing for LuaJIT.
  19. 19.
    Documentation - Our Documentation.
  20. 20.
    NumPy — NumPy.

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Faculty of Automatic Control and ComputersUniversity Politehnica of BucharestBucharestRomania

Personalised recommendations