🚀 go-pugleaf

In article <8134@sirius.ucs.adelaide.edu.au> snoswell@ucs.adelaide.edu.au (Michael Snoswell) writes:
>I recently read an article on maximum entropy image processing (S.F
>Gull and J. Skilling, IEE Proceedings, Vol 131, Pt.F, No. 6, OCTOBER
>1984) and thought it was something that I could use instead on 2D
>deconvolution.

Maximum entropy is basically just a particular slant on 2-D convolution
that takes into account certain properties of the image and noise.  It's
definitely not a computational shortcut.

>I'm not a real whiz at maths so the salient details were over my head.
>However, I was wondering if there is a C implementation available via
>anonymous FTP and if the site details could be posted or mailed to me?

I saw this recently in this newsgroup:

>From: mathew@elroy.jpl.nasa.gov (Mathew Yeates)
>I have written a Khoros Toolbox which contains several image restoration
>routines. Before you can use it, however, you need to obtain Khoros.
>Khoros is available through anon ftp at pprg.eece.unm.edu.
>See also the newsgroup comp.soft-sys.khoros

Hope that helps.


--
Stan Reeves
Auburn University, Department of Electrical Engineering, Auburn, AL  36849
INTERNET: sjreeves@eng.auburn.edu

John Skilling has a company called Maximum Entropy Data Consultants, Ltd. that
markets his MEM code (called MEMSYS5).  This is generally considered the state-
of-the-art MEM package, at least within the astronomy image processing community.
I know they sell to a lot of medical imaging people too.  If you are interested,
MEDC can be contacted at 20 Cockroft Place, Cambridge, England  CB3 0HF.  
Tel. (44)223-312562.  Send e-mail to mark@vax.ra.physics.cambridge.ac.uk.

There are some PD packages too, but the ones I know of are layered on larger
astronomical data reduction and analysis packages.  NRAO's AIPS software includes
an MEM implementation, and NOAO's IRAF package just recently has a package.
The AIPS software is written in Fortran, and IRAF software is written in a
pre-processed language similar to RATFOR.
---
Robert J. Hanisch                       Tel. (410)338-4910   Fax. (410)338-5090
Space Telescope Science Institute       NSI: hanisch@stsci.edu [130.167.1.2]
3700 San Martin Drive                   DECNet: STSCIC::HANISCH [6549::] 
Baltimore, MD  21218  USA               Bitnet: hanisch@stsci

In article <sjreeves.920811141812@fourier.ee.eng.auburn.edu> sjreeves@eng.auburn.edu (Stan Reeves) writes:
>In article <8134@sirius.ucs.adelaide.edu.au> snoswell@ucs.adelaide.edu.au (Michael Snoswell) writes:
>>I recently read an article on maximum entropy image processing (S.F
>>Gull and J. Skilling, IEE Proceedings, Vol 131, Pt.F, No. 6, OCTOBER
>>1984) and thought it was something that I could use instead on 2D
>>deconvolution.
>
>Maximum entropy is basically just a particular slant on 2-D convolution
                                                         ^^^^^^^^^^^^^^^
>that takes into account certain properties of the image and noise.  It's
>definitely not a computational shortcut.

Oops!  I meant 2-D DEconvolution.


--
Stan Reeves
Auburn University, Department of Electrical Engineering, Auburn, AL  36849
INTERNET: sjreeves@eng.auburn.edu

In article <1992Aug12.153110.15062@stsci.edu> hanisch@stsci.edu (Bob Hanisch,N406,4910) writes:

> John Skilling has a company called Maximum Entropy Data Consultants, Ltd. that
> markets his MEM code (called MEMSYS5).  This is generally considered the state-
> of-the-art MEM package, at least within the astronomy image processing community.

While MEM -- and particularly the MEDC package -- seems to be pretty
effective for astronomical imagery, I'd like to raise a small word of
warning regarding MEM for general restoration applications.  MEM works
by taking a uniform image and `pulling' non-uniformities out of it
only where there is statistical evidence for doing so.  Unfortunately,
the staistical models used by all the implementations of MEM that I've
seen only work well for pulling bright regions from a dark
backgrouond.  Given a image with the opposite contract (e.g., a
bright-field electron micrograph) I've yet to see a good job.

If anyone can:

   - take an image
   - work out the MEM restoration
   - invert the contrast of the image
   - work out the MEM reconstruction of the contrast-inverted image
   - invert the contrast of this reconstruction

and get similar results I, for one, would be most interested!

--
 Dr Adrian F. Clark                                   JANET: alien@uk.ac.essex
 INTERNET: alien%uk.ac.essex@nsfnet-relay.ac.uk          FAX: (+44) 206-872900
 BITNET: alien%uk.ac.essex@ac.uk              PHONE: (+44) 206-872432 (direct)
 Dept ESE, University of Essex, Wivenhoe Park, Colchester, Essex, C04 3SQ, UK.

sjreeves@eng.auburn.edu (Stan Reeves) writes:

>In article <8134@sirius.ucs.adelaide.edu.au> snoswell@ucs.adelaide.edu.au (Michael Snoswell) writes:
>>I recently read an article on maximum entropy image processing (S.F
>>Gull and J. Skilling, IEE Proceedings, Vol 131, Pt.F, No. 6, OCTOBER
>>1984) and thought it was something that I could use instead on 2D
>>deconvolution.

>>From: mathew@elroy.jpl.nasa.gov (Mathew Yeates)
>>I have written a Khoros Toolbox which contains several image restoration

There is indeed an implementation of Gulls "intrinsic correlation"
scheme in the toolbox. See "Maximum entropy and Bayesian methods,
Cambridge 1988".

Unfortunately, the rules are changing here regarding the release of
code. I'm going to have to jump through some hoops before I can release
version 0.2 but I'm still hopeful. Since version 0.2 will come with
readymade "workspaces" it will be much easier to use. But since 
you were interested in MaxEnt, version 0.1 might work for you.
It's available in the contrib directory of Khoros.

-Mathew

In article <ALIEN.92Aug13121817@vulcan.essex.ac.uk> alien@essex.ac.uk ("Adrian F Clark") writes:
>While MEM -- and particularly the MEDC package -- seems to be pretty
>effective for astronomical imagery, I'd like to raise a small word of
>warning regarding MEM for general restoration applications.  MEM works
>by taking a uniform image and `pulling' non-uniformities out of it
>only where there is statistical evidence for doing so.  Unfortunately,
>the staistical models used by all the implementations of MEM that I've
>seen only work well for pulling bright regions from a dark
>backgrouond.  Given a image with the opposite contract (e.g., a
>bright-field electron micrograph) I've yet to see a good job.

The results from MaxEnt are only as good as your prior assumptions. In
astronomy, it is reasonable to say 'my prior expectation of what my ccd
measures is typically 10 photons/s, but may be from 1e-4 to 1e5 photons/s'.
This is obviously an assymetric distribution, because of the knowledge of
a minimum signal level. (i.e. there are no negative photons out there). It's
not invariant under inverting the contrast, and rightly so.

If you want such invariance, then it's easy to attain; either throw away
the positivity constraint if you don't know what the zero level is, or
introduce a maximum level if you have upper and lower bounds for you image.

The MEDC code supports both of those.


David Robinson. (drtr1@cus.cam.ac.uk)

I would like to collect some evidence that image processing is moving more
and more into the mainstream of consumer electronics.  More specifically,
I am trying to find information that would motivate a greater emphasis on
teaching image processing to undergraduates in electrical engineering (if,
indeed, this is justified).

Examples that I've thought of are:

- HDTV will soon move into the mainstream and will require engineers with
  a working knowledge of image and video signal processing.
- Increasingly sophisticated VCR's and camcorders are beginning to use
  image processing.
- AT&T's picture phone points to the fact that true videophones are just
  around the corner.
- The possibility of telecomputers that provide random access to TV programs
  and interactivity --- this will soon be feasible with continued growth in
  available bandwidth and transmission technology.

Any thoughts along these lines would be greatly appreciated.  I will summarize
email responses to the net.



--
Stan Reeves
Auburn University, Department of Electrical Engineering, Auburn, AL  36849
INTERNET: sjreeves@eng.auburn.edu