Re: image rotation
- From: Phil Carmody <thefatphil_demunged@xxxxxxxxxxx>
- Date: 25 Nov 2005 19:30:27 +0200
"Nishu" <naresh_earth@xxxxxxxxx> writes:
> yes, flipping diagonally is obviously tough in rectanglar cases,
> otherwise it would have been cakewalk.
>
> Doing Pixel by pixel mapping using the concept of discompostion in
> disjoint cycles and then implementing the algo seems the best solution
> as suggested by some professor. still i'm facing the problem of
> flagging the traversed pixels. I may have to lose some precision to get
> the thing working. :(
With no extra memory, you can simulate a run through the cycle, and if
you encounter an index smaller than where you started, then you've
already done the cycle, so don't do it for real. I have no idea what
the asymptotics of this might be.
Phil
--
If a religion is defined to be a system of ideas that contains unprovable
statements, then Godel taught us that mathematics is not only a religion, it
is the only religion that can prove itself to be one. -- John Barrow
.
- Follow-Ups:
- Re: image rotation
- From: megagurka
- Re: image rotation
- References:
- image rotation
- From: Nishu
- Re: image rotation
- From: Thomas Richter
- Re: image rotation
- From: Willem
- Re: image rotation
- From: Nishu
- image rotation
- Prev by Date: Re: What is the state-of-the-art analysing hardware impact on achievable compression ratio
- Next by Date: arith coding patents
- Previous by thread: Re: image rotation
- Next by thread: Re: image rotation
- Index(es):
Relevant Pages
|
|