Hope this helps,

From what I understand the down-sample is different from the wavelet. Read
how to do a Harr Wavelet encoding, without the number of lifts stored you
can't recover to original signal. A Harr Wavelet  is really basic algebra,
but a very counterintuitive algorithm.

jungle.war....@gmail.com wrote:
> IS it possible operation???

Well no, if I've got it right, you want to recover what you did away
with during downsampling, without using the components resulting from
the result of your DWT on your original signal, right?

Dear ALL!!
Suppose, I have image (256x256):
img = im2double(imread('lena.bmp'));

Then I perform discrete wavelet transform and get 4 components
of source image:
[a, h, v, d] = dwt2(img, 'db1');

Then I reconstruct signal using only detail components:

det = idwt2([], h, v, d, 'db1');

Next I downsample det with 3/4, i.e. get
det2 with size 192x192:

det2 = downsample(det, 3/4); - not matlab code


And I want to know can i get det from det2:
det = reconstruct(det2);

What filters i must use to perform such operation???
IS it possible operation???