An Experimental Comparison of Three Guiding Principles for the Detection of Salient Image Locations: Stability, Complexity, and Discrimination

  • Dashan Gao
  • Nuno Vasconcelos
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4840)


We present an experimental comparison of the performance of representative saliency detectors from three guiding principles for the detection of salient image locations: locations of maximum stability with respect to image transformations, locations of greatest image complexity, and most discriminant locations. It is shown that discriminant saliency performs better in terms of 1) capturing relevant information for classification, 2) being more robust to image clutter, and 3) exhibiting greater stability to image transformations associated with variations of 3D object pose. We then investigate the dependence of discriminant saliency on the underlying set of candidate discriminant features, by comparing the performance achieved with three popular feature sets: the discrete cosine transform, a Gabor, and a Haar wavelet decomposition. It is show that, even though different feature sets produce equivalent results, there may be advantages in considering features explicitly learned from examples of the image classes of interest.


Discrete Cosine Transform Image Class Saliency Detector Image Transformation Interest Point Detector 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Harris, C., Stephens, M.: A combined corner and edge detector. In: Alvey Vision Conference, pp. 147–151 (1988)Google Scholar
  2. 2.
    Förstner, W.: A framework for low level feature ex-traction. In: Eklundh, J.-O. (ed.) ECCV 1994. LNCS, vol. 800, pp. 383–394. Springer, Heidelberg (1994)CrossRefGoogle Scholar
  3. 3.
    Sha’ashua, A., Ullman, S.: Structural saliency: the detection of globally salient structures using a locally connected network. In: Proc. ICCV, pp. 321–327 (1988)Google Scholar
  4. 4.
    Lindeberg, T.: Scale-space theory: A basic tool for analyzing structures at different scales. J. Applied Statistics 21(2), 224–270 (1994)Google Scholar
  5. 5.
    Mikolajczyk, K., Schmid, C.: Indexing based on scale invariant interest points. In: Proc. ICCV, pp. 525–531 (2001)Google Scholar
  6. 6.
    Mikolajczyk, K., Schmid, C.: An affine invariant interest point detector. In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2350, pp. 128–142. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  7. 7.
    Lowe, D.G.: Object recognition from local scale-invariant features. In: Proc. ICCV, pp. 1150–1157 (1999)Google Scholar
  8. 8.
    Sebe, N., Lew, M.S.: Comparing salient point detectors. Pattern Recognition Letters 24(1-3), 89–96 (2003)CrossRefzbMATHGoogle Scholar
  9. 9.
    Kadir, T., Brady, M.: Scale, saliency and image description. Int’l. J. Comp. Vis. 45, 83–105 (2001)CrossRefzbMATHGoogle Scholar
  10. 10.
    Itti, L., Koch, C., Niebur, E.: A model of saliency-based visual attention for rapid scene analysis. IEEE Trans. PAMI 20(11) (1998)Google Scholar
  11. 11.
    Privitera, C., Stark, L.: Algorithms for defining visual regions-of-interest: comparison with eye fixations. IEEE Trans. PAMI 22, 970–982 (2000)CrossRefGoogle Scholar
  12. 12.
    Gao, D., Vasconcelos, N.: Discriminant saliency for visual recognition from cluttered scenes. In: Proc. NIPS, pp. 481–488 (2004)Google Scholar
  13. 13.
    Walker, K., Cootes, T., Taylor, C.: Locating salient object features. In: Proc. British Machine Vision Conf., pp. 557–566 (1998)Google Scholar
  14. 14.
    Schiele, B., Crowley, J.: Where to look next and what to look for. In: Intelligent Robots and Systems (IROS), pp. 1249–1255 (1996)Google Scholar
  15. 15.
    Burt, P., Adelson, E.H.: The laplacian pyramid as a compact image code. IEEE Trans. Communication COM-31, 532–540 (1983)CrossRefGoogle Scholar
  16. 16.
    Schmid, C., Mohr, R., Bauckhage, C.: Evaluation of interest point detectors. Int’l J. Comp. Vis. 37(2), 151–172 (2000)CrossRefzbMATHGoogle Scholar
  17. 17.
    Fergus, R., Perona, P., Zisserman, A.: Object class recognition by unsupervised scale-invariant learning. In: Proc. IEEE Conf. CVPR (2003)Google Scholar
  18. 18.
    Kadir, T., Zisserman, A., Brady, M.: An affine invariant saliency region detector. In: Pajdla, T., Matas, J(G.) (eds.) ECCV 2004. LNCS, vol. 3021, pp. 228–241. Springer, Heidelberg (2004)CrossRefGoogle Scholar
  19. 19.
    Nene, S., Nayar, S., Murase, H.: Columbia object image library: Coil-100. Technical Report CUCS-006-96, Dept. of Computer Science, Columbia Univ. (1996)Google Scholar
  20. 20.
    Vasconcelos, N., Carneiro, G.: What is the role of independence for visual regognition? In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 2002. LNCS, vol. 2350, Springer, Heidelberg (2002)Google Scholar
  21. 21.
    Daugman, J.: Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimensional visual cortical filters. J. Opt. Soc. Am. A 2(7), 1362–1373 (1985)CrossRefGoogle Scholar
  22. 22.
    Viola, P., Jones, M.: Robust real-time object detection. In: 2nd Int. Workshop on Statistical and Computational Theories of Vision Modeling, Learning, Computing and Sampling (July 2001)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Dashan Gao
    • 1
  • Nuno Vasconcelos
    • 1
  1. 1.Statistical Visual Computing Laboratory, Department of Electrical and Computer Engineering, University of California San Diego, La Jolla, CA 92093USA

Personalised recommendations