Newsgroups: sci.image.processing
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!europa.eng.gtefsd.com!howland.reston.ans.net!EU.net!Germany.EU.net!unibwh.unibw-hamburg.de!unibwh!e_meier
From: e_meier@unibw-hamburg.de (Wilhelm Meier)
Subject: Estimation of pdf
Message-ID: <E_MEIER.94Nov3131139@siegen.unibw-hamburg.de>
Sender: news@unibw-hamburg.de
Reply-To: e_meier@unibw-hamburg.de
Organization: University of the Federal Armed Forces, Hamburg
Date: Thu, 3 Nov 1994 12:11:39 GMT
Lines: 34


I'm not very skilled in the field of statistics, so don't be surprised
if this is a very simple problem:

I have a histogram of some image data. I know that there are N different, 
statistical indepent sources of gaussian type, with different means,
variances and different a-priori probabilities. These 3N quantities
are unknown. I want to estimate these parameters using the data of
the histogram. How can I formulate an effective estimation procedure.
Actually, I've tested a maximization algorithm (Newton-like), which
directly maximizes the Likelihood. For N=1 or N=2 it works quite well,
but for N>=3 it takes a lot of time, since the parameter space is 
3N-dimensional. 
Is there a method, which does the estimation of the parameters
WITHOUT iteration - something like a moment-based approach ?
If there exists an effective method, is it applicable to non-gaussian
types of distribution function - like chi-square, etc.

Thanks for all comments !
I'll summarize - mail directly !
 
Wilhelm


--
                             _\\|//_ 
                             ( O-O )
---------------------------o00--(_)--00o------------------------------
Dipl.-Ing. Wilhelm Meier	Phone:	(49)(0)40 / 65412524
UniBwH, FB ET			FAX:	(49)(0)40 / 65412822
Holstenhofweg 85		email:	e_meier@unibw-hamburg.de
D-22043	Hamburg, GERMANY


