|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[WM]: Image distorsioned inserting a watermark usign Cox algorithm: MORE DETAILS
[WM]: Image distorsioned inserting a watermark usign Cox et = al.
algorithm
-----------------------------------------------------------------------
Hi,
I am trying to insert a watermark into a image using the = Cox et al.
algorithm described into "Secure Spread Spectrum watermarkin= g for multimedia", but I have some problems.
I apply the dct on entire image, only for luminance compo= nent (using a dct matrix version, with a reasonable cost for little
images) and then I find the N greatest dct coefficients to i= nsert the watermark(generated randomly with a N(0,1) distribution) by
= the equation vi'=vi*(1+ alpha*xi). Then, my problem is that I have obse= rved the original image is lightly and visible
modified.
For more details:
* I use N=1000 (number of dct coefficients to be marked)
* I use alpha=0.1
* I select the N greatest coefficients from the AC coe= fficients. I don't use the DC coefficient.
Anybody can give me some idea about my problem?
To select the N largest coefficients, I must consider th= e sign or the value without sign?
Has anybody obtained the same results or similar that th= ose described into the Cox's paper?? I'll be grateful for any aportation.
Thanks for advanced.
tpj
______________________________________________________________________________
Watermarking Mailing List - http://www.watermarkingworld.org/ml.html
To unsubscribe send email to "majordomo@watermarkingworld.org" with
"unsubscribe watermarking YOURMAIL" in the body.
______________________________________________________________________________
|