An Approach to the Compact and Efficient Visual Codebook Based on SIFT Descriptor

  • Zhe Wang
  • Guizhong Liu
  • Xueming Qian
  • Danping Guo
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6297)


The Bag-of-Words (BoW) derived from local keypoints was widely applied in visual information research such as image search, video retrieval, object categorization, and computer vision. Construction of visual codebook is a well-known and predominant method for the representation of BoW. However, a visual codebook usually has a high dimension that results in high computational complexity. In this paper, an approach is presented for constructing a compact visual codebook. Two important parameters, namely the likelihood ratio and the significant level, are proposed to estimate the discriminative capability of each of the codewords. Thus, the codewords that have higher discriminative capability are reserved, and the others are removed. Experiments prove that application of the proposed compact codebook not only reduces computational complexity, but also improves performance of object classification..


SIFT descriptor visual codebook compact codebook 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Jiang, Y., Ngo, C., Yang, J.: Towards optimal bag-of-features for object categorization and semantic video retrieval. In: CIVR (2007)Google Scholar
  2. 2.
    Mikolajczyk, K., Schmid, C.: A performance evaluation of local descriptors. In: CVPR (2003)Google Scholar
  3. 3.
    Lowe, D.: Distinctive image features from scale invariant keypoints. IJCV 60(2), 91–110 (2004)CrossRefGoogle Scholar
  4. 4.
    Li, F.F., Fergus, R., Perona, P.: Learning generative visual models from few training examples: an incremental bayesian approach tested on 101 object categories. In: CVPR Workshop on Generative-Model Based Vision (2004)Google Scholar
  5. 5.
    Kim, S., Kweon, I.S., Lee, C.W.: Visual Categorization Robust to Large Intra-Class Variations using Entropy-guided Codebook. In: IEEE International Conference on Robotics and Automation Roma, Italy, April 10-14 (2007)Google Scholar
  6. 6.
    Wu, L.N., Luo, S.W., Sun, W.: Create efficient visual codebook based on weighted mRMR for object categorization. In: ICSP 9th International Conference (2008)Google Scholar
  7. 7.
    Hotta, K.: Object Categorization Based on Kernel Principal Component Analysis of Visual Words. In: WACV (2008)Google Scholar
  8. 8.
    Chang, S.F., He, J.F., et al.: Columbia University/VIREO-CityU/IRIT TRECVID2008 High-Level Feature Extraction and Interactive Video Search. In: Proc. TRECVID 2008 (2008)Google Scholar
  9. 9.
    Snoek, C.G.M., Van, K.E.A., et al.: The MediaMill TRECVID 2008 Semantic Video Search Engine. In: Proc. TRECVID 2008 (2008)Google Scholar
  10. 10.
    Wang, L.: Toward a discriminative codebook: codeword selection across multi-resolution. In: IEEE Conference on Computer Vision and Pattern Recognition, June 17-22, pp. 1–8 (2007)Google Scholar
  11. 11.
    Li, T., Mei, T., Kweon, I.S.: Learning Optimal Compact Codebook for Efficient Object Categorization. In: WACV (2008)Google Scholar
  12. 12.
    Chang, C.C., Lin, C.J.: LIBSVM: a library for support vector machines (2001),
  13. 13.
    TREC Video Retrieval Evaluation (TRECVID),
  14. 14.
    Everingham, M., Zisserman, A., et al.: The 2005 pascal visual object classes challenge. In: Quiñonero-Candela, J., Dagan, I., Magnini, B., d’Alché-Buc, F. (eds.) MLCW 2005. LNCS (LNAI), vol. 3944, pp. 117–176. Springer, Heidelberg (2006)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Zhe Wang
    • 1
  • Guizhong Liu
    • 1
  • Xueming Qian
    • 1
  • Danping Guo
    • 1
  1. 1.School of Electronics and Information EngineeringXi’an Jiaotong UniversityChina

Personalised recommendations