Prev: 32bit vs. 64bit, MEX, and OS X 10.6
Next: Efficient Frontier Weights
From: Omar on 7 May 2010 23:18
In a project I am working on I need to map an image to a feature space, so as to reduce the dimensions. I have to do it by "selected the 3000 strongest wavelet coefficients of the image at level 3" (assume 'db1' wavelet).
Can anyone help me with how to go about it? Once processing is done in the feature space, how should I map back?
From: Wayne King on 8 May 2010 07:28
"Omar " <omar.nadeem(a)hotmail.com> wrote in message <hs2l5c$ppn$1(a)fred.mathworks.com>...
> In a project I am working on I need to map an image to a feature space, so as to reduce the dimensions. I have to do it by "selected the 3000 strongest wavelet coefficients of the image at level 3" (assume 'db1' wavelet).
> Can anyone help me with how to go about it? Once processing is done in the feature space, how should I map back?

Hi Omar, Do you mean you want to retain the largest wavelet coefficients at level 3 regardless of whether they are vertical, horizontal, or diagonal details. In other words, just the largest coefficients at level 3 irrespective of which details they are?

Wayne
 | 
Pages: 1
Prev: 32bit vs. 64bit, MEX, and OS X 10.6
Next: Efficient Frontier Weights