sbanner1@uvicctr.UUCP (S. John Banner) (07/22/87)
A few days ago I posted a message asking for algorithms that will
efficently thin out a bitmap, and I got a message back asking for
clarification. So, in case there are other people out there who
didn't understand what I was asking before, I will try to explain
it again, and be a bit more clear this time. If this sounds like
an ammazingly simple question, please humor me, I don't do a lot
of graphics work (or at least I havn't up till now), and my first
graphics course isn't till next spring.
The problem:
To take a bitmap, that may contain areas that have all the pixels
turned on, and convert it to a bitmap where no two pixels within
some given radius are turned on, while retaining as much of the
original as possible.
For example, if I have the following picture, and want a radius of
1 (ie. no pixel that is turned on, is adjacent to any other pixel that
is also turned on), I should get something like:
... ... . . . .
... ...
... ... \ . . . .
......... _____\ .
......... > . .
......... _____/ . .
... /
... . .
...
Note that this is a very simple and contrived example, but I think
it gets the idea accross. The major problem I see with the algorithms
that I have heard of, is that if you have a dithered image (I am working
in Black, and White, but a future enhancement MAY include color), giving
you a grey scale, you will come out with an image that has basicly had
the contrast turned just about to the maximum.
As before, any help what so ever will be greatly appreciated,
and I will post a summary, unless I am told not to,
S. John Banner
...!uw-beaver!uvicctr!sol!sbanner1
...!ubc-vision!uvicctr!sol!sbanner1
ccsjb@uvvm
sbanner1@sol.UVIC.CDN