Re: Accurate edge detection?

> >  It would be very simple to set an appropriate threshold and cut
> > all the rest of diffraction off, but I didn't do it especially to show
> > restoration technique. I've just have restored some text instead of my
> > first set of objects
(http://www.smartfills.com/Html/Images/textdiffr.jpg
> > andhttp://www.smartfills.com/Html/Images/textrestr.jpg)
I do not
> > think that there is a way (and necessity) make it better.
>
> Spoken like a true resarcher. Where is it published?
>
> illywhacker;

In font of your eyes, only. The last sentence of your own caused my
strong desire to tell you some stories about researchers and why this
word sounds like mating grunt for me. But, this is obviously wrong
place for it. If you have an interest in technique what I've used to
revert irreversible transform, you'd better contact me directly. I
see, that all another boys lost interest in this discussion, and as
for me I unfortunately have some problems when I have to load each
time 600K of traffic what is a must to get the last message in this
topic. Nothing to do, I live in the backside of the civilized world
and my Net is limited to 15- rarely 20 Kbits (not kilobytes!) per sec.
I found this Google group three weeks ago and realized that this one
has some interest for me. I know both physic, math and programming-
all components what are a must to solve such sort of problems like
restoration of diffracted images. So I spent two days, playing with
algorithms. I tested three variants of how it could be done, and later
found my own old book on computers in image processing. At this book I
found description of one method, what I had tested and declined
finally. My variant works better, but not considerably better. The
only thing what I have discovered (and what was a real surprise for
me!) was that relatively simple methods can produce good results. I
mean, that what I used was the only linear algebra- no linear
programming.
mats(at)smartfills(dot)com