Adaptative Markov Random Fields for Omnidirectional Vision

Images obtained with catadioptric sensors contain significant deformations which prevent the direct use of classical image treatments. Thus, Markov random fields (MRF) whose usefulness is now obvious for projective image processing, cannot be used directly on catadioptric images because of the inadequacy of the neighborhood. In this paper, we propose to define a new neighborhood for MRF by using the equivalence theorem developed for central catadioptric sensors. We show the importance of this adaptation for a motion detection application

[1]  G. B. Smith,et al.  Preface to S. Geman and D. Geman, “Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images” , 1987 .

[2]  P. Pérez,et al.  Multiscale minimization of global energy functions in some visual recovery problems , 1994 .

[3]  A. Makadia,et al.  Image processing in catadioptric planes: spatiotemporal derivatives and optical flow computation , 2002, Proceedings of the IEEE Workshop on Omnidirectional Vision 2002. Held in conjunction with ECCV'02.

[4]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Kostas Daniilidis,et al.  Catadioptric Projective Geometry , 2001, International Journal of Computer Vision.

[6]  Shree K. Nayar,et al.  A Theory of Single-Viewpoint Catadioptric Image Formation , 1999, International Journal of Computer Vision.

[7]  J. Besag.  Spatial Interaction and the Statistical Analysis of Lattice Systems , 1974 .

[8]  Helder Araújo,et al.  Direct Least Square Fitting of Paracatadioptric Line Images , 2003, 2003 Conference on Computer Vision and Pattern Recognition Workshop.

[9]  Helder Araújo,et al.  Pose estimation for central catadioptric systems: an analytical approach , 2002, Object recognition supported by user interaction for service robots.