Neural Network based Novelty Detection for Incremental Semi-supervised Learning in Multi-class Gesture Recognition

Husam Al-Behadili, Arne Grumpe, Christian Wöhler

2016

Abstract

The problems of infinitely long data streams and its concept drift as well as non-linearly separable classes and the possible emergence of “novel classes” are topics of high relevance for gesture data streaming based automatic recognition systems. To address these problems we apply a semi-supervised learning technique using a neural network in combination with an incremental update rule. Neural networks have been shown to handle non-linearly separable data and the incremental update ensures that the parameters of the classifier follow the “concept-drift” without the necessity of an increased training set. Since a semi-supervised learning technique is sensitive to false labels, we apply an outlier detection method based on extreme value theory and confidence band intervals. The proposed algorithm uses the extreme learning machine, which is easily updated and works with multi-classes. A comparison with an auto-encoder neural network shows that the proposed algorithm has superior properties. Especially, the processing time is greatly reduced.

References

  1. Al-Behadili, H., Wöhler, C., and Grumpe, A. (2014). Semi-supervised learning of emblematic gestures. AtAutomatisierungstechnik, 62(10):732-739.
  2. Al-Behadili, H., Wöhler, C., and Grumpe, A. (2015). Extreme learning machine based novelty detection for incremental semi-supervised learning. In 3'rd International conference on image Information Processing (ICIIP), page In Press. IEEE.
  3. Clifton, D. A., Clifton, L. A., Bannister, P. R., and Tarassenko, L. (2008). Automated novelty detection in industrial systems. In Advances of Computational Intelligence in Industrial Systems, pages 269- 296. Springer.
  4. Fisher, R. A. and Tippett, L. H. C. (1928). Limiting forms of the frequency distribution of the largest or smallest member of a sample. In Mathematical Proceedings of the Cambridge Philosophical Society, volume 24, pages 180-190. Cambridge Univ Press.
  5. Hertz, J., Krogh, A., and Palmer, R. G. (1991). Introduction to the theory of neural computation, volume 1. Basic Books.
  6. Huang, G., Song, S., Gupta, J., and Wu, C. (2014). Semi-supervised and unsupervised extreme learning machines. IEEE Transactions on Cybernetics, 44(12):2405-2417.
  7. Huang, G.-B., Chen, L., and Siew, C.-K. (2006a). Universal approximation using incremental constructive feedforward networks with random hidden nodes. IEEE Transactions on Neural Networks, 17(4):879-892.
  8. Huang, G.-B., Zhu, Q.-Y., and Siew, C.-K. (2006b). Extreme learning machine: theory and applications. Neurocomputing, 70(1):489-501.
  9. Hugueny, S., Clifton, D. A., and Tarassenko, L. (2009). Novelty detection with multivariate extreme value theory, part ii: An analytical approach to unimodal estimation. In Proc. MLSP, pages 1-6. IEEE.
  10. Johnson, R. and Zhang, T. (2015). Semi-supervised learning with multi-view embedding: Theory and application with convolutional neural networks. Proc. CoRR.
  11. Kardaun, O. J. (2005). Classical methods of statistics: with applications in fusion-oriented plasma physics, volume 1. Springer Science & Business Media.
  12. Kendall, W. S., Marin, J.-M., and Robert, C. P. (2007). Confidence bands for brownian motion and applications to monte carlo simulation. Statistics and Computing, 17(1):1-10.
  13. Lan, Y., Soh, Y. C., and Huang, G.-B. (2009). Ensemble of online sequential extreme learning machine. Neurocomputing, 72(13):3391-3395.
  14. Li, K., Zhang, J., Xu, H., Luo, S., and Li, H. (2013). A semisupervised extreme learning machine method based on co-training. Journal of Computational Information Systems, 9(1):207-214.
  15. Masud, M. M., Gao, J., Khan, L., Han, J., and Thuraisingham, B. (2011). Classification and novel class detection in concept-drifting data streams under time constraints. IEEE Transactions on Knowledge and Data Engineering, 23(6):859-874.
  16. Pimentel, M., Clifton, D., Clifton, L., and Tarassenko, L. (2014). A review of novelty detection. Signal Processing, 99:215-249.
  17. Rao, C. R. and Mitra, S. K. (1971). Generalized inverse of matrices and its applications, volume 7. Wiley New York.
  18. Richarz, J. and Fink, G. A. (2011). Visual recognition of 3d emblematic gestures in an hmm framework. Journal of Ambient Intelligence and Smart Environments, 3(3):193-211.
  19. Roberts, S. J. (1999). Novelty detection using extreme value statistics. IEE Proceedings-Vision, Image and Signal Processing, 146(3):124-129.
  20. Sakic, D. (2012). Semi-supervised learning using ensemble methods gestures recognition. Master's thesis, University of Dortmund.
  21. Tax, D. (2001). One-class classification: concept-learning in the absence of counter-examples. PhD thesis, TU Delft, Delft University of Technology.
  22. Tax, D. (2015). Ddtools, the data description toolbox for matlab.
  23. Tax, D. M. J. and Duin, R. P. W. (1999). Support vector domain description. Pattern Recognition Letters, 20(11- 13):1191-1199.
  24. Yusoff, Y. A., Basori, A. H., and Mohamed, F. (2013). Interactive hand and arm gesture control for 2d medical image and 3d volumetric medical visualization. Procedia-Social and Behavioral Sciences, 97:723- 729.
  25. Zhu, X. and Goldberg, A. B. (2009). Introduction to semisupervised learning. Synthesis lectures on artificial intelligence and machine learning, 3(1):1-130.
Download


Paper Citation


in Harvard Style

Al-Behadili H., Grumpe A. and Wöhler C. (2016). Neural Network based Novelty Detection for Incremental Semi-supervised Learning in Multi-class Gesture Recognition . In Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP, (VISIGRAPP 2016) ISBN 978-989-758-175-5, pages 287-294. DOI: 10.5220/0005674202870294


in Bibtex Style

@conference{visapp16,
author={Husam Al-Behadili and Arne Grumpe and Christian Wöhler},
title={Neural Network based Novelty Detection for Incremental Semi-supervised Learning in Multi-class Gesture Recognition},
booktitle={Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP, (VISIGRAPP 2016)},
year={2016},
pages={287-294},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005674202870294},
isbn={978-989-758-175-5},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 11th Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP, (VISIGRAPP 2016)
TI - Neural Network based Novelty Detection for Incremental Semi-supervised Learning in Multi-class Gesture Recognition
SN - 978-989-758-175-5
AU - Al-Behadili H.
AU - Grumpe A.
AU - Wöhler C.
PY - 2016
SP - 287
EP - 294
DO - 10.5220/0005674202870294