-----Original Message-----
    From: Steve Niemela <[log in to unmask]>
    Newsgroups: bit.listserv.confocal
    To: [log in to unmask] <[log in to unmask]>
    Date: Friday, March 24, 2000 3:57 PM
    Subject: Deconvolution software
    
    
    > Second message:
    > 
    > To clear up some of the recommendations surrounding deconvolution
    > software, I think the following information should be noted.
    > 
    > There are four kinds of deconvolution algorithms.
    > 
    > 1. "Nearest Neighbor" - will run very fast on even the junkiest
    Ø computer,  however, does not result in much improvement
    
    Again, this is really off.  The nearest neighbor, if it is implemented
    correctly and if it uses a good PSF is very valuable.  In fact, the
    differences between images deconvolved with VayTek's nearest neighbor
    and a constrained iterative and a blind deconvolution, is small.  If you
    place these images side by side you can see the differences, but they
    are not big differences.  You can solve a lot of problems using a
    nearest neighbor algorithm.  This is a myth promulgated by those
    companies that have a particular axe to grind.
    
    

Steve
 
sorry but you can't be serious with this statement. The nearest neighbour algorithm has almost no mathematical foundation. However I agree with you that there still is use for NN for quick cleanup but not for high quality restorations. It makes me wonder though, why, as you say, the differences to a constrained iterative method are so small on your system.
 
The NN attempts to reduce the 3D problem into a series of 2D calculations. The neighbouring
focal sections on either side of the in-focus section are taken as an approximation to the real image structure either side. Their re-blurred (with the respective PSF slice) images are assumed to be an estimate of additional out-of-focus light added. A fraction of this estimate is then subtracted by the in-focus plane. ( D. AGARD, Optical sectioning microscopy: cellular architecture in three dimensions, Biophysics Bioengineering 1984, No 13, pp. 191-219 )

Further, with neither the nearest neighbour nor the inverse filter it is possible to increase the resolution of a microscope. 


Cheers
Lutz

______________________________________
Lutz Schaefer
Advanced Imaging Methodology Consultation
16-715 Doon Village Rd.
Kitchener, Ontario
N2P 2A2, Canada
Email: [log in to unmask]
Phone, FAX: (519)-894-8870
______________________________________