Peleg, S.[Shmuel],
A New Probabilistic Relaxation Scheme,
PAMI(2), No. 4, July 1980, pp. 362-369.
BibRef
8007
Earlier:
PRIP79(337-343).
BibRef
Earlier:
A Note on the Advantages of a New Probabilistic Relaxation Scheme,
UMD-TR-739, February 1979.
A relaxation updating functions is derived using probability theory.
The primary difference is that it is multiplicative (not additive)
(for combining Q values). A method for deriving for compatibility
function is also given (this seem more useful for image type data
rather than graph matching).
BibRef
Peleg, S.[Shmuel], and
Rosenfeld, A.[Azriel],
A Note on the Evaluation of Probabilistic Labelings,
SMC(11), No. 2, February 1981, pp. 176-179.
Relaxation, Evaluation. This is an initial report on evaluation of probabilistic methods
used in relaxation. There are no firm conclusions reached about the
various evaluation methods.
BibRef
8102
Peleg, S.[Shmuel], and
Rosenfeld, A.[Azriel],
Breaking Substitution Ciphers Using a Relaxation Algorithm,
CACM(22), No. 11, November 1979, pp. 598-605.
BibRef
7911
Shvaytser, H.[Haimd], and
Peleg, S.,
A New Approach to the Continuous Labeling Problem,
Hebrew Univ. of Jerusalem in
CVPR85(320-327).
A new formulation of the equations and what is
computed. Requires some study to determine what it really does.
BibRef
8500
O'Leary, D.P.,
Peleg, S.[Shmuel],
Analysis of Relaxation Processes: The Two-Node Two-Label Case,
SMC(13), 1983, pp. 618-623.
BibRef
8300
Peleg, S.[Shmuel],
Labeling Evaluation in Probabilistic Networks,
InfoSys(21), 1980, pp. 213-220.
BibRef
8000
Peleg, S.[Shmuel],
Monitoring Relaxation Algorithms Using Labeling Evaluations,
ICPR80(54-57).
(Maryland).
More relaxation theory.. How to find a proper stopping point.
BibRef
8003
Chapter on Matching and Recognition Using Volumes, High Level Vision Techniques, Invariants continues in
Faugeras and Berthod Gradient Optimization Methods .