Abstract
Deterministic annealing and relaxation labeling algorithms for classification and matching are presented and discussed. A new approach —self annealing—is introduced to bring deterministic annealing and relaxation labeling into accord. Self annealing results in an emergent linear schedule for winner-take-all and assignment problems. Also, the relaxation labeling algorithm can be seen as an approximation to the self annealing algorithm for matching and labeling problems.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
J. S. Bridle. Training stochastic model recognition algorithms as networks can lead to maximum mutual information estimation of parameters. In D. S. Touretzky, editor, Advances in Neural Information Processing Systems 2, pages 211–217, San Mateo, CA, 1990. Morgan Kaufmann.
J. Buhmann and T. Hofmann. Central and pairwise data clustering by competitive neural networks. In J. Cowan, G. Tesauro, and J. Alspector, editors, Advances in Neural Information Processing Systems 6, pages 104–111. Morgan Kaufmann, San Francisco, CA, 1994.
W. J. Christmas, J. Kittler, and M. Petrou. Structural matching in computer vision using probabilistic relaxation. IEEE Trans. Patt. Anal. Mach. Intell., 17(5):749–764, Aug. 1995.
R. Duda and P. Hart. Pattern Classification and Scene Analysis. Wiley, New York, NY, 1973.
O. Faugeras and M. Berthod. Improving consistency and reducing ambiguity in stochastic labeling: an optimization approach. IEEE Trans. Patt. Anal. Mach. Intell., 3(4):412–424, Jul. 1981.
A. H. Gee and R. W. Prager. Polyhedral combinatorics and neural networks. Neural Computation, 6(1):161–180, Jan. 1994.
D. Geiger and A. L. Yuille. A common framework for image segmentation. Intl. Journal of Computer Vision, 6(3):227–243, Aug. 1991.
S. Gold and A. Rangarajan. A graduated assignment algorithm for graph matching. IEEE Transactions on Pattern Analysis and Machine Intelligence, 18(4):377–388, 1996.
E. R. Hancock and J. Kittler. Discrete relaxation. Pattern Recognition, 23(7):711–733, 1990.
J. J. Hopfield and D. Tank. 'Neural’ computation of decisions in optimization problems. Biological Cybernetics, 52:141–152, 1985.
R. Hummel and S. Zucker. On the foundations of relaxation labeling processes. IEEE Trans. Patt. Anal. Mach. Intell., 5(3):267–287, May 1983.
A. K. Jain and R. C. Dubes. Algorithms for Clustering Data. Prentice Hall, Englewood Cliffs, NJ, 1988.
B. Kamgar-Parsi and B. Kamgar-Parsi. On problem solving with Hopfield networks. Biological Cybernetics, 62:415–423, 1990.
J. Kivinen and M. K. Warmuth. Exponentiated gradient versus gradient descent for linear predictors. Technical Report UCSC-CRL-94-16, Univ. Calif. Santa Cruz, June 1994.
J. J. Kosowsky and A. L. Yuille. The invisible hand algorithm: Solving the assignment problem with statistical physics. Neural Networks, 7(3):477–490, 1994.
M.Pelillo. On the dynamics of relaxation labeling processes. In IEEE Intl. Conf. on Neural Networks (ICNN), volume 2, pages 606–1294. IEEE Press, 1994.
M. Pelillo. Learning compatibility coefficients for relaxation labeling processes. IEEE Trans. Patt. Anal. Mach. Intell, 16(9):933–945, Sept. 1994.
C. Peterson and B. Söderberg. A new method for mapping optimization problems onto neural networks. Intl. Journal of Neural Systems, 1(1):3–22, 1989.
A. Rangarajan, S. Gold, and E. Mjolsness. A novel optimizing network architecture with applications. Neural Computation, 8(5):1041–1060, 1996.
A. Rangarajan, A. L. Yuille, S. Gold, and E. Mjolsness. A convergence proof for the softassign quadratic assignment algorithm. In Advances in Neural Information Processing Systems (NIPS) 9. MIT Press, 1997. (in press).
A. Rosenfeld, R. Hummel, and S. Zucker. Scene labeling by relaxation operations. IEEE Trans. Syst. Man, Cybern., 6(6):420–433, Jun. 1976.
R. Sinkhorn. A relationship between arbitrary positive matrices and doubly stochastic matrices. Ann. Math. Statist., 35:876–879, 1964.
F. R. Waugh and R. M. Westervelt. Analog neural networks with local competition. I. Dynamics and stability. Physical Review E, 47(6):4524–4536, June 1993.
A. L. Yuille and J. J. Kosowsky. Statistical physics algorithms that converge. Neural Computation, 6(3):341–356, May 1994.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Rangarajan, A. (1997). Self annealing: Unifying deterministic annealing and relaxation labeling. In: Pelillo, M., Hancock, E.R. (eds) Energy Minimization Methods in Computer Vision and Pattern Recognition. EMMCVPR 1997. Lecture Notes in Computer Science, vol 1223. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-62909-2_83
Download citation
DOI: https://doi.org/10.1007/3-540-62909-2_83
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-62909-2
Online ISBN: 978-3-540-69042-9
eBook Packages: Springer Book Archive