next up previous
Next: Maximum Likelihood Estimate Up: Parameter Estimation for Previous: Parameter Estimation for

Least Squares Estimate of Parameters

 

The least squares estimate detailed in [3] assumes that the observations of the GMRF image obey the model

where is a zero mean correlated noise sequence with variance and correlation with the following structure:

 

Then, for , the LSE are:

where is the complete set of pixels, and toroidal wrap-around is assumed.



next up previous
Next: Maximum Likelihood Estimate Up: Parameter Estimation for Previous: Parameter Estimation for



David A. Bader
dbader@umiacs.umd.edu