Combining geometric and probabilistic structure for active recognition of 3D objects
Direct perception is incomplete: objects may show ambiguous appearances, and sensors have a limited sensitivity. Consequently, the recognition of complex 3D objects necessitates an exploratory phase to be able to deal with complex scenes or objects.
The variation of object appearance when the viewpoint is modified or when the sensor parameters are changed is an idiosyncratic feature which can be organized in the form of an aspect graph.
Standard geometric aspect graphs are difficult to build. This article presents a generalized probabilistic version of this concept. When fitted with a Markov chain dependance, the aspect graph acquires a quantitative predictive power. Tri-dimensional object recognition becomes translated into a problem of Markov chain discrimination. The asymptotic theory of hypothesis testing, in its relation to the theory of large deviations, gives then a global evaluation of the statistical complexity of the recognition problem.
Keywords3D object recognition active vision aspect graphs Markov chains statistical hypothesis testing
Unable to display preview. Download preview PDF.
- 1.T. Arbel and F.P. Ferrie. Informative views and sequential recognition. In Proc. Fourth European Conference on Computer Vision, Cambridge, England, April 14–18, pages I:469–481. 1996.Google Scholar
- 3.P. Billingsley. Statistical Inference for Markov Processes. The University of Chicago Press, Chicago, London, 1961.Google Scholar
- 6.A. Dembo and O. Zeitouni. Large Deviations Techniques and Applications. Jones and Bartlett Publishers, Boston, 1993.Google Scholar
- 7.S.J. Dickinson, H.I. Christensen, J. Tsotsos, and G. Olofsson. Active object recognition integrating attention and viewpoint control. In J-O. Eklundh, editor, Proc. Third European Conference on Computer Vision, Stockholm, Sweden, May 2–6, number 801 in Lecture Notes in Computer Science, pages 3–14. Springer Verlag, Berlin, 1994.Google Scholar
- 12.M.R. Garey and D.S. Johnson. Computers and Intractability — A guide to the theory of NP-completeness. W.H. Freeman, San Francisco, 1979.Google Scholar
- 13.Y. Gdalyahu and D. Weinshall. Measures for silhouettes resemblance and representative silhouettes of curved objects. In B.F. Buxton and R. Cipolla, editors, Proc. Fourth European Conference on Computer Vision, Cambridge, England, April 14–18, volume 1065 of Lecture Notes in Computer Science, pages 363–375. Springer Verlag, Berlin, Heidelberg, New York, 1996.Google Scholar
- 14.S. Herbin. Recognizing 3D objects by generating random actions. In Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, CA, June 18–20, pages 35–40, 1996.Google Scholar
- 15.S. Herbin. Eléments pour la formalisation d'une reconnaissance active — application à la vision tri-dimensionnelle. Thèse de mathématiques appliquées, Ecole Normale Supérieure de Cachan, Juillet 1997. In french.Google Scholar
- 17.J. Köbler, U. Schöning, and J. Torn. The Graph Isomorphism Problem: Its Structural Complexity. Birkhauser, Cambridge, MA, 1993.Google Scholar
- 18.J. Koenderink. Solid Shape. MIT Press, Cambridge Ma, 1990.Google Scholar
- 20.T. Lindeberg. Scale-space theory: A basic tool for analysing structures at different scales. Journal of Applied Statistics, 21(2):224–270, 1994.Google Scholar
- 21.T. Lindeberg. Scale-Space Theory in Computer Vision. 1994.Google Scholar
- 26.J.H. Rieger. On the complexity and computation of view graphs of piecewise-smooth algebraic surfaces. Technical Report FBI_HH_M_228/93, Hamburg UniveritÄt, 1993.Google Scholar
- 29.E Seneta. Non-negative matrices and Markov chains. Springer Verlag, New York, second edition, 1981.Google Scholar
- 30.L.G. Shapiro and M.S. Costa. Appearance-based 3D object recognition. In M. Hebert, J. Ponce, T.E. Boult, A. Gross, and D. Forsyth, editors, Proc. International Workshop on 3-D Object Representation in Computer Vision, New York City, NY, USA, 5–7 Dec 1994, volume 994 of Lecture Notes in Computer Science, pages 51–64. Springer Verlag, Berlin, 1995.Google Scholar
- 31.I. Shimshoni and J. Ponce. Probabilistic 3D object recognition. In Proc. Fifth International Conference on Computer Vision, Cambridge, MA, USA, June 20–22, pages 488–493, 1995.Google Scholar
- 35.T. van Effelterre. Calcul exact du graphe d'aspect de solides de révolution. Thèse de doctorat, Université de Rennes I, 1995.Google Scholar
- 36.S. Vinther and R. Cipolla. Active 3D object recognition using 3D affine invariants. In J-O. Eklundh, editor, Proc. Third European Conference on Computer Vision, Stockholm, Sweden, May 2–6, number 801 in Lecture Notes in Computer Science, pages 15–24. Springer Verlag, Berlin, 1994.Google Scholar
- 37.D. Weinshall, M. Werman, and N. Tishby. Stability and likelihood of views of three dimensional objects. In J-O. Eklundh, editor, Proc. Third European Conference on Computer Vision, Stockholm, Sweden, May 2–6, number 800 in Lecture Notes in Computer Science, pages 24–35. Springer Verlag, Berlin, 1994.Google Scholar
- 38.D. Wilkes and J.K. Tsotsos. Active object recognition. In Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Champaign, IL, June 15–18, pages 136–141, 1992.Google Scholar