Neural Network Based Eye Tracking

  • Pavel Morozkin
  • Marc Swynghedauw
  • Maria Trocan
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10449)


The EyeDee embedded eye tracking solution developed by SuriCog is the world’s first solution using the eye as a real-time mobile digital cursor, while maintaining full mobility. In order to reduce the time of eye image transmission, image compression techniques can be employed. Being hardware implemented, several standard image coding systems (JPEG and JPEG2000) were evaluated for their potential use in the next generation device of the EyeDee product line. In order to satisfy low-power, low-heat, low-MIPS requirements several non-typical approaches have been considered. One example consists in the complete replacement of currently used eye tracking algorithm based on image processing coupled with geometric eye modeling by a precisely tuned and perfectly trained neural network, which directly transforms wirelessly transmitted floating-point values of decimated eye image (result of the 3D perspective projection of a model of rotating pupil disk) into five floating-point parameters of pupil’s ellipse (result of the eye tracking). Hence implementation of the eye tracking algorithm is reduced to a known challenge of neural network construction and training, preliminary results of which are presented in the paper.


Eye tracking Human–machine interaction Neural networks 


  1. 1.
    Morozkin, P., Swynghedauw, M., Trocan, M.: Design of an embedded image acquisition system. In: 2015 IEEE International Conference on Electronics, Circuits, and Systems (ICECS), pp. 504–505 (2015)Google Scholar
  2. 2.
    Morozkin, P., Swynghedauw, M., Trocan, M.: An image compression for embedded eye-tracking applications. In: 2016 International Symposium on INnovations in Intelligent SysTems and Applications (INISTA), article 33 (2016)Google Scholar
  3. 3.
    Morozkin, P., Swynghedauw, M., Trocan, M.: Image quality impact for eye tracking systems accuracy. In: 2016 IEEE International Conference on Electronics, Circuits and Systems (ICECS), pp. 429–431 (2016)Google Scholar
  4. 4.
    Antonini, M., Barlaud, M., Mathieu, P., Daubechies, I.: Image coding using wavelet transform. IEEE Trans. Image Process. 1(2), 205–220 (1992)CrossRefGoogle Scholar
  5. 5.
    Demuth, H.B., Beale, M.H., De Jesus, O., Hagan, M.T.: Neural Network Design. Martin Hagan, Oklahoma (2014)Google Scholar
  6. 6.
    Torch framework. Accessed 17 May 2017
  7. 7.
    Vogl, T.P., Mangis, J.K., Rigler, A.K., Zink, W.T., Alkon, D.L.: Accelerating the convergence of the back-propagation method. Biol. Cybern. 59(4), 257–263 (1988)CrossRefGoogle Scholar
  8. 8.
    Mason, L., Baxter, J., Bartlett, P.L., Frean, M.R.: Boosting algorithms as gradient descent. In: NIPS, pp. 512–518 (1999)Google Scholar
  9. 9.
    Ng, A.Y.: Feature selection, L1 vs. L2 regularization, and rotational invariance. In: Proceedings of the Twenty-First International Conference on Machine Learning, p. 78. ACM (2004)Google Scholar
  10. 10.
    Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction, vol. 1. MIT Press, Cambridge (1998)Google Scholar
  11. 11.
    Srivastava, N., Hinton, G.E., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15(1), 1929–1958 (2014)MathSciNetzbMATHGoogle Scholar
  12. 12.
    Ioffe, S., Szegedy, C.: Batch normalization: accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 (2015)
  13. 13.
    Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems, pp. 1097–1105 (2012)Google Scholar
  14. 14.
    He, K., Zhang, X., Ren, S., Sun, J.: Delving deep into rectifiers: surpassing human-level performance on imagenet classification. In: Proceedings of the IEEE International Conference on Computer Vision, pp. 1026–1034 (2015)Google Scholar
  15. 15.
    Xu, B., Wang, N., Chen, T., Li, M.: Empirical evaluation of rectified activations in convolutional network. arXiv preprint arXiv:1505.00853 (2015)

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Pavel Morozkin
    • 1
    • 2
  • Marc Swynghedauw
    • 1
  • Maria Trocan
    • 2
  1. 1.SuriCogParisFrance
  2. 2.Institut Supérieur d’Electronique de ParisParisFrance

Personalised recommendations