Contents
Markov random fields
 
X is a Markov random field :
the value of a pixel of X only depends on its neighbours
 
 
A
 
 
 
 
 
 
 
 
 
D
S
B
 
 
 
 
 
 
 
 
 
C
 
        
First order Neighbours  and corresponding cliques
{A,B,C,D} = neighbourhood of S

Hammersley-Clifford theorem :

X is a Markov random field 
X follows a Gibbs distribution.
U(X) is an energy defined with respect to the cliques
 
 
André Jalobeanu - 24 Aug 1998