Improving Accuracy for Image Parsing Using Spatial Context and Mutual Information

  • Thi Ly Vu
  • Sun-Wook Choi
  • Chong Ho Lee
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8228)


This paper presents a novel approach for image parsing based on nonparametric model in superpixel level. Spatial context and mutual information between object co-occurrence are introduced and applied for improving the accuracy of image parsing. These methods make the probability of object co-occurrence more reliable, and thus the inference of object label from K nearest neighbors is more accurate. Our system integrates the probability of object co-occurrence with the spatial context and mutual information into a Markov Random Field(MRF) framework. Experimental results on SIFTFlow and Barcelona dataset shows that the spatial context and the mutual information are promising methods to improve the accuracy of nonparametric image parsing models.


image parsing MRF superpixel spatial context mutual information SIFTFlow 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Galleguillos, C., Rabinovich, A., Belongie, S.: Object categorization using Co-Occurrence, Location and Appearance. In: CVPR (2008)Google Scholar
  2. 2.
    Choi, M.J., Lim, J.J., Torralba, A., Willsky, A.S.: Exploiting Hierarchical context on a large database of object categories. In: CVPR (2010)Google Scholar
  3. 3.
    Felzenszwalb, P.F., Huttenlocher, D.P.: Efficient Graph-Based Image Segmentation. IJCV (2003)Google Scholar
  4. 4.
    Tighe, J., Lazebnik, S.: SuperParsing: Scalable Nonparametric Image parsing with super-pixels. In: CVPR (2010)Google Scholar
  5. 5.
    Eigen, D., Fergus, R.: Nonparametric image parsing using adaptive neighbor sets. In: CVPR (2012)Google Scholar
  6. 6.
    Myeong, H., Chang, J.Y., Lee, K.M.: Learning object relationships via graph-based context model. In: CVPR (2012)Google Scholar
  7. 7.
    Chow, C.K., Liu, C.N.: Approximating discrete probability distributions with dependence tree. In: IEEE (1968)Google Scholar
  8. 8.
    He, X., Zemel, R.S., Carreira-Perpinan, M.A.: Multiscale conditional random fields for image labeling. In: CVPR (2004)Google Scholar
  9. 9.
    Shotton, J., Winn, J., Rother, C., Criminisi, A.: Textonboost for image understanding: Multiclass object recognition and segmentation by jointly modeling appearance, shape and context. IJCV (2007)Google Scholar
  10. 10.
    Singhal, A., Luo, J., Zhu, W.: Probabilistic spatial context models for scene content understanding. In: CVPR (2003)Google Scholar
  11. 11.
    Church, K.W., Hanks, P.: Words association norms, mutual information and lexicography. In: Proceedings of the 27th Annual Conference of the Association for Computational Linguistics (1989)Google Scholar
  12. 12.
    Tighe, J., Lazebnik, S.: SuperParsing: Scalable Nonparametric Image parsing with super-pixels. IJCV (2013)Google Scholar
  13. 13.
    Kumar, S., Hebert, M.: A hierarchical field framework for unified context-based classification. In: ICCV (2005)Google Scholar
  14. 14.
    Liu, C., Lim, J., Torralba, A.: Nonparametric scene parsing: Label transfer via dense scene alignment. In: IEEE (2011)Google Scholar
  15. 15.
    Tighe, J., Lazebnik, S.: Understanding scenes on many levels. In: CVPR (2011)Google Scholar
  16. 16.
    Matsuo, Y., Ishizuka, M.: Keyword extraction from a single document using word co-occurrence statistical information. International Journal on Artificial Intelligence Tools (2003)Google Scholar
  17. 17.
    Russell, B.C., Torralba, A., Fergus, R., Freeman, W.T.: Object regconition by scene alignment. In: NIPS (2007)Google Scholar
  18. 18.
    Weeds, J., Weir, D.: Co-occurrence retrieval: A flexible framework for lexical sistributional similarity. In: Computational Linguistic (2006)Google Scholar
  19. 19.
    Dataset for image parsing experiments,

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Thi Ly Vu
    • 1
  • Sun-Wook Choi
    • 1
  • Chong Ho Lee
    • 1
  1. 1.School of Information and Communication EngineeringInha UniversityIncheonKorea

Personalised recommendations