Skip to main content
SpringerLink
Log in

Condensing Deep Fisher Vectors: To Choose or to Compress?

  • Conference paper
  • First Online:
Pattern Recognition Applications and Methods (ICPRAM 2017)

Part of the book series: Lecture Notes in Computer Science ((LNIP,volume 10857))

Abstract

Feature selection and dimensionality reduction are the two popular off-the-shelf techniques in practice for reducing data’s high dimensional memory footprint and thus making it amenable for large scale visual retrieval and classification. In this paper, we show that feature compression is a better choice than feature selection when dealing with large scale retrieval of high dimensional Fisher vectors derived from deep or shallow stochastic models such as restricted Boltzmann machine (RBM). The dimensionality of the Fisher vectors is proportional to the size of the architecture from which they are drawn. As the number of hidden units in RBM increases, the dimensionality of the Fisher vectors also scales accordingly, thus increasing storage requirements as well as causing overfitting during classification. In order to tackle these challenges, we compare the performance of feature compression and feature selection techniques and suggest the use of compression methods on available Fisher encodings. We have based our diagnostics on multi-collinearity evaluation metrics and justify the use of the proposed feature condensation method using feature visualisations and classification accuracy on benchmark data set.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Log in via an institution
Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 39.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 54.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

References

  1. Csurka, G., Dance, C., Fan, L., Willamowski, J., Bray, C.: Visual categorization with bags of keypoints. In: Workshop on Statistical Learning in Computer Vision, ECCV, Prague (2004)

    Google Scholar 

  2. Everingham, M., Van Gool, L., Williams, C.K., Winn, J., Zisserman, A.: The PASCAL visual object classes (VOC) challenge. Int. J. Comput. Vis. 88(2), 303–338 (2010)

    Article  Google Scholar 

  3. Griffin, G., Holub, A., Perona, P.: Caltech-256 object category dataset. California Institute of Technology, Technical report, 7694 (2007). http://authors.library.caltech.edu/7694

  4. Fei-Fei, L., Fergus, R., Perona, P.: One-shot learning of object categories. IEEE Trans. Pattern Anal. Mach. Intell. 28(4), 594–611 (2006)

    Article  Google Scholar 

  5. Farquhar, J., Szedmak, S., Meng, H., Shawe-Taylor, J.: Improving “bag-of-keypoints” image categorisation: generative models and PDF-kernels. Univ. Southampton 68 (2005)

    Google Scholar 

  6. Perronnin, F., Dance, C., Csurka, G., Bressan, M.: Adapted vocabularies for generic visual categorization. In: Leonardis, A., Bischof, H., Pinz, A. (eds.) ECCV 2006. LNCS, vol. 3954, pp. 464–475. Springer, Heidelberg (2006). https://doi.org/10.1007/11744085_36

    Chapter  Google Scholar 

  7. Boureau, Y., Bach, F., LeCun, Y., Ponce, J.: Learning mid-level features for recognition. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition (2010)

    Google Scholar 

  8. Wang, G., Hoiem, D., Forsyth, D.: Learning image similarity from Flickr groups using stochastic intersection kernel machines. In: 2009 IEEE 12th International Conference on Computer Vision. IEEE (2009)

    Google Scholar 

  9. Lazebnik, S., Schmid, C., Ponce, J.: Beyond bags of features: spatial pyramid matching for recognizing natural scene categories. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2006). IEEE (2006)

    Google Scholar 

  10. Jaakkola, T., Haussler, D.: Exploiting generative models in discriminative classifiers. In: Advances in Neural Information Processing Systems, vol. 11, pp. 487–493. MIT Press (1998)

    Google Scholar 

  11. Perronnin, F., Dance, C.: Fisher kernels on visual vocabularies for image categorization. In: IEEE Conference on Computer Vision and Pattern Recognition. IEEE (2007)

    Google Scholar 

  12. Perronnin, F., Sánchez, J., Mensink, T.: Improving the fisher kernel for large-scale image classification. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010. LNCS, vol. 6314, pp. 143–156. Springer, Heidelberg (2010). https://doi.org/10.1007/978-3-642-15561-1_11

    Chapter  Google Scholar 

  13. Chatfield, K., Lempitsky, V., Vedaldi, A., Zisserman, A.: The devil is in the details: an evaluation of recent feature encoding methods. In: Proceedings of the British Machine Vision Conference. BMVA Press (2011)

    Google Scholar 

  14. Perronnin, F., Larlus, D.: Fisher vectors meet neural networks: a hybrid classification architecture. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, CVPR (2015)

    Google Scholar 

  15. Sanchez, J., Perronnin, F., Mensink, T., Verbeek, J.: Compressed fisher vectors for large-scale image classification. Rapport de recherche RR-8209, INRIA, January 2013

    Google Scholar 

  16. Zhang, Y., Wu, J., Cai, J.: Compact representation for image classification: to choose or to compress? In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society (2014)

    Google Scholar 

  17. Ahmed, S., Azim, T.: Compression techniques for deep fisher vectors. In: Proceedings of the 6th International Conference on Pattern Recognition Applications and Methods (ICPRAM) (2017)

    Google Scholar 

  18. Azim, T., Niranjan, M.: Inducing discrimination in biologically inspired models of visual scene recognition. In: IEEE International Workshop on Machine Learning for Signal Processing (MLSP) (2013)

    Google Scholar 

  19. Azim, T.: Visual scene recognition with biologically relevant generative models. Ph.D. thesis, School of Electronics and Computer Science, Southampton, UK (2014)

    Google Scholar 

  20. Maaten, L.: Learning discriminative fisher kernels. In: Proceedings of the 28th International Conference on Machine Learning, ICML 2011, Bellevue, Washington, USA, 28 June–2 July 2011, pp. 217–224. Omnipress (2011)

    Google Scholar 

  21. Hinton, G.: Training products of experts by minimizing contrastive divergence. Neural Comput. 14, 1771–1800 (2002)

    Article  Google Scholar 

  22. Hinton, G.E.: A practical guide to training restricted Boltzmann machines. In: Montavon, G., Orr, G.B., Müller, K.-R. (eds.) Neural Networks: Tricks of the Trade. LNCS, vol. 7700, pp. 599–619. Springer, Heidelberg (2012). https://doi.org/10.1007/978-3-642-35289-8_32

    Chapter  Google Scholar 

  23. Jayalakshmi, T., Santhakumaran, A.: Statistical normalization and back propagation for classification. Int. J. Comput. Theor. Eng. 3(1), 89 (2011)

    Article  Google Scholar 

  24. Weiss, Y., Torralba, A., Fergus, R.: Spectral hashing. In: Advances in Neural Information Processing Systems, NIPS (2009)

    Google Scholar 

  25. Hinton, G., Salakhutdinov, R.: Reducing the dimensionality of data with neural networks. Science 313, 504–507 (2006)

    Article  MathSciNet  Google Scholar 

  26. Maaten, L.: Learning a parametric embedding by preserving local structure. In: RBM (2009)

    Google Scholar 

  27. Fleuret, F.: Fast binary feature selection with conditional mutual information. J. Mach. Learn. Res. 5, 1531–1555 (2004)

    MathSciNet  MATH  Google Scholar 

  28. Peng, H., Long, F., Ding, C.: Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy. IEEE Trans. Pattern Anal. Mach. Intell. 27(8), 1226–1238 (2005)

    Article  Google Scholar 

  29. Guyon, I., Weston, J., Barnhill, S., Vapnik, V.: Gene selection for cancer classification using support vector machines. Mach. Learn. 46(1), 389–422 (2002)

    Article  Google Scholar 

  30. Breiman, L.: Random forests. Mach. Learn. 45(1), 5–32 (2001)

    Article  Google Scholar 

  31. Belsley, D.A.: A guide to using the collinearity diagnostics. Comput. Econ. 4(1), 33–50 (1991)

    MathSciNet  MATH  Google Scholar 

  32. LeCun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proc. IEEE 86(11), 2278–2324 (1998)

    Article  Google Scholar 

  33. Bottou, L., Bousquet, O., Zurich, G.: The tradeoffs of large scale learning. In: Advances in Neural Information Processing Systems, pp. 161–168 (2008)

    Google Scholar 

  34. Everingham, M., Eslami, S., Van Gool, L., Williams, C., Winn, J., Zisserman, A.: The PASCAL visual object classes challenge: a retrospective. Int. J. Comput. Vis. 111(1), 98–136 (2015)

    Article  Google Scholar 

  35. Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Sean, M., Zhiheng, H., Karpathy, A., Khosla, A., Bernstein, M., Berg, A., Fei-Fei, L.: ImageNet large scale visual recognition challenge. Int. J. Comput. Vis. (IJCV) 115(3), 211–252 (2015)

    Article  MathSciNet  Google Scholar 

Download references

Acknowledgement

This research (SRGP: 21-402) was supported by Higher Education Commission (HEC) of Pakistan & NVIDIA (Ref.: 281400) with a valuable donation of Titan-X graphics card.

Author information

Authors and Affiliations

  1. Center of Excellence in IT, Institute of Management Sciences, Peshawar, Pakistan

    Sarah Ahmed & Tayyaba Azim

Authors
  1. Sarah Ahmed
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Tayyaba Azim
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Sarah Ahmed or Tayyaba Azim .

Editor information

Editors and Affiliations

  1. Sapienza Università di Roma, Rome, Italy

    Maria De Marsico

  2. ICAR-CNR, Naples, Napoli, Italy

    Gabriella Sanniti di Baja

  3. University of Lisbon, Lisbon, Portugal

    Ana Fred

Rights and permissions

Reprints and permissions

Copyright information

© 2018 Springer International Publishing AG, part of Springer Nature

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Ahmed, S., Azim, T. (2018). Condensing Deep Fisher Vectors: To Choose or to Compress?. In: De Marsico, M., di Baja, G., Fred, A. (eds) Pattern Recognition Applications and Methods. ICPRAM 2017. Lecture Notes in Computer Science(), vol 10857. Springer, Cham. https://doi.org/10.1007/978-3-319-93647-5_5

Download citation

  • DOI: https://doi.org/10.1007/978-3-319-93647-5_5

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-319-93646-8

  • Online ISBN: 978-3-319-93647-5

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics

Search

Navigation