Towards Vision-Based Understanding of Unknown Environments

  • A. Śluzek
  • M. Paradowski
Part of the Advances in Intelligent and Soft Computing book series (AINSC, volume 98)


The paper demonstrates how to transform (using a combination of techniques reported in our previous papers) a collection of random images gathered in an unknown environment into a limited-scale visual model of that environment. The model generally consists of the template images of the typical “visual objects” identified in the explored world. Both the concepts of objects and their templates are formed without any assumptions about the content of acquired images, i.e. the semantics is built using the pictorial data only (although users may subsequently identify the real-world semantics of the formed objects). From the image processing perspective, the method consists in detecting near-duplicate (i.e. photometric/geometric distortions and partial occlusions are allowed) fragments in random images. It is envisaged that such a proposal can be instrumental in assisting both autonomous agents and visually impaired humans (including both blind people and people unable to understand perceived visual data) facing unfamiliar worlds. The paper focuses on the practical aspects of the problem (exemplary results, computational efficiency, etc.) although a substantial amount of theoretical background is also included.


Autonomous Agent Visual Object Query Image Affine Transformation Unknown Environment 
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. [Biederman 1987]
    Biederman, I.: Recognition-by-components: A theory of human image understanding. Psychological Review 94(2), 115–147 (1987)CrossRefGoogle Scholar
  2. [Bourbakis and Kavraki 2001]
    Bourbakis, N.G., Kavraki, D.: An intelligent assistant for navigation of visually impaired people. In: 2nd IEEE Int. Symp. on Bioinformatics and Bioengineering, Bethesda, p. 230 (2001)Google Scholar
  3. [Ke et al. 2004]
    Ke, Y., Sukthankar, R., Huston, L.: Efficient near-duplicate detection and sub-image retrieval. In: ACM Multimedia Conference, New York, pp. 869–876 (2004)Google Scholar
  4. [Matas et al. 2002]
    Matas, J., Chum, O., Urban, M., Pajdla, T.: Robust wide baseline stereo from maximally stable extremal regions. In: British Machine Vision Conf., Cardiff, pp. 384–393 (2002)Google Scholar
  5. [Mikolajczyk and Schmid 2004]
    Mikolajczyk, K., Schmid, C.: Scale and affine invariant interest point detectors. Int. J. of Computer Vision 60, 63–86 (2004)CrossRefGoogle Scholar
  6. [Paradowski and Śluzek 2010]
    Paradowski, M., Śluzek, A.: Local keypoints and global affine geometry: Triangles and ellipses for image fragment matching. In: Kwaśnicka, H., Jain, L.C. (eds.) Innovations in Intelligent Image Analysis. SCI, vol. 339, pp. 195–224. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  7. [Paradowski and Śluzek 2010a]
    Paradowski, M., Śluzek, A.: Automatic visual object formation using image fragment matching. In: 5th Int. Symp. Advances in Artificial Intelligence & Applications, Wisła, pp. 97–104 (2010)Google Scholar
  8. [Riesenhuber and Poggio 2000]
    Riesenhuber, M., Poggio, T.: Models of object recognition. Nature Neuroscience 3, 1199–1204 (2000)CrossRefGoogle Scholar
  9. [Zhao and Ngo 2009]
    Zhao, W.-L., Ngo, C.-W.: Scale-rotation invariant pattern entropy for keypoint-based near-duplicate detection. IEEE Trans. on Image Processing 18(2), 412–423 (2009)MathSciNetCrossRefGoogle Scholar
  10. [Zhao et al. 2007]
    Zhao, W.-L., Ngo, C.-W., Tan, H.-K., Wu, X.: Near-duplicate keyframe identification with interest point matching and pattern learning. IEEE Trans. on Multimedia 9(5), 1037–1048 (2007)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • A. Śluzek
    • 1
    • 2
  • M. Paradowski
    • 3
  1. 1.Nanyang Technological UniversitySingapore
  2. 2.Nicolaus Copernicus UniversityToruńPoland
  3. 3.Wrocław University of TechnologyPoland

Personalised recommendations