How MEM works
There is usually no unique inverse to the process of converting
x-rays into a light curve for a Rotating Modulation Collimator
(RMC) like RHESSI. If there were, it would be possible to
minimize the chisquare (or other statistic) that measures the
goodness of fit between the observed and modeled count rates.
One of the ways around this is to reduce the infinite number of
possible solutions to the one that has the maximum entropy:
We want to minimize chisquare and maximize entropy. This is impossible
so we maximize the quantity
where 
is a free parameter of the reconstruction.
The MEM_SATO program uses the same technique as the program of the same
name used for Yohkoh/HXT maps.
RETURN
Ed Schmahl
Last modified: Sun Apr 7 22:01:52 EDT 2002