Re: rotation by 90 in only one buffer
- From: Hugo Pfoertner <nothing@xxxxxxxxxxxx>
- Date: Tue, 29 Nov 2005 01:13:12 +0100
Nishu wrote:
>
> Hii all,
>
> what could be the BEST Memory Optimal way of rotating an image or
> matrix by 90 degrees using input buffer only? i.e. Input buffer only is
> the output memory buffer. so that we have parameters as (inputbuffer,
> width, height of 2D array)
>
> Say, we can use small amount of memory location for temporary storage
> of any block/blocks in the image/array but not the whole of
> image/array.
>
> rotation of sqaure matrix is comparitively simple but i'm facing
> problem with rectanglar matrix.
>
> Thanks and Regards,
> Naresh
Some links and counts for small matrices are given in Sloane's OEIS:
http://www.research.att.com/projects/OEIS?Anum=A093055
http://www.research.att.com/projects/OEIS?Anum=A093056
http://www.research.att.com/projects/OEIS?Anum=A093057
Unfortunately the link to P.F. Windley's article "Transposing Matrices
in a Digital Computer" given in A093055 is outdated. ->
http://comjnl.oxfordjournals.org/cgi/content/abstract/2/1/47
A discussion of available methods is given in Knuth's TAOCP, Vol. 1,
page 523.
Hugo
.
- References:
- rotation by 90 in only one buffer
- From: Nishu
- rotation by 90 in only one buffer
- Prev by Date: Re: particle triangle collision
- Next by Date: Re: Sphere silhouette, circle parametrization
- Previous by thread: Re: rotation by 90 in only one buffer
- Next by thread: dobashi clouds
- Index(es):
Relevant Pages
|