Re: [WM]: coefficient in LL subband after 2-D DWT haar =

["Amir Tavala" <amir.iut@gmail.com>]

 Dear Jian Yi

   Well, it is not strange at all. The approximate coefficients of a 8-bit *m*n *image contains some pixels which their values are
larger than 255 cause of the haar formula which puts the mean value of a 2*2 block as the value of LL coefficients, So the size
would be *m/2 * n/2*.
   If there's still any problem I'll be glad to help.

Regards
Amir M. Tavala

>
> On 10/4/06, =1B$B7z0l=1B(B(Jian Yi) <im943084@mail.cpu.edu.tw> wrote:
> >
> > Dears adviser
> > I have a question about the image coefficient in LL subband after 
> > 2-d=20 DWT transformaton.
> > I have a gray image , its bit depth is 256.
> > Because I try to implement a DWT transformation in MATLAB, I find 
> > the=20 coefficients in the LL subband are always larger 255.
> > I use the 'haar' filter, so I think it's ok.
> > But, I always study some paper where their experimetn result, LL 
> > band=20 come  DWT haar transform looks the same as the 1/4 origional 
> > image. I=20 dont know why ? the pixels is larger than 255??
> > Just like the attachement.
> >
> >
> > I wonder Why?
> >
> > Is there some concept I misunderstand?
> > Is there some parameter needed to be used in MATLAB ?
> >
> > Please help me.
> >
> > Thanks your answers for me.
> >
> >
> > Wishes
> >
> >
> > Jian-Yi
> >
>
> --
> Amir M. Tavala



______________________________________________________________________________

Watermarking Mailing List - http://www.watermarkingworld.org/ml.html
To unsubscribe send email to "majordomo@watermarkingworld.org" with
"unsubscribe watermarking YOURMAIL" in the body.
______________________________________________________________________________