Re: [WM]: buliding the digital watermarking software

[Stefan Katzenbeisser <e9625414@student.tuwien.ac.at>]

Hi,

nuenhua muegrabeesungharn wrote:
> 
>        I have tested two softwares. They are JTEG on Windows
> (non-commercial) and  Steganos Suite II ( commercial and has just released)
> . JSTEG seems to be built based on  LSB modification  which  I simply tested
> by changing the size of the image . 

No. JSTEG uses the quantization of DCT coefficients to transmit information.

> The result was the digital watermark
> couldn't  be recovered. The Steganos Suite II,it can hide the data only into
> the bmp format( for image) :( no JPEG .I read the description .This program
> seems to use the LSB modification method -_-' too. 

I don't really know what the Steganos suite uses as hiding  method. It is very 
likely that it's some sort of LSB modification, though. Both methods are not 
suitable for watermarking, as you found out yourself.

> I couldnt test it because
> it performed the illegal operation on my machine (laptop 266mmx) .

It's always nice to see these blue screens ;)

> I got a question on my mind. When they want to use the LSB method do they need to
> convert the compressed image into the uncompressed image (I'm not sure about
> this) such as to convert the .jpg to tiff/targa format then modify the LSB
> of 8 bit pixel . 

LSB modification normally only works in the spatial domain with a lossless
compression system (JPEG is lossy and works in the DCT domain). So you have
to convert JPEG files to some lossless format (like BMP). Unfortunately, if
you want ro re-compress the result, it is very likely that you won't be 
able to recover the watermark.

>         I wonder if there's any available software that the digital
> watermark system is based on the spread spectrum ,dct modification, to
> transform the image into the frequency domain  .

Yes: http://www.cl.cam.ac.uk/~fapp2/watermarking/products.html

>         Hm. and if the algorithm is based on the sptread spectrum .Do we
> need the original image to extract the watermark ? 

Depends. There are methods which require the orininal image in the extraction
process and there are methods which don't. 

>          I intend to do the digital watermark on the jpeg. thus, I guess I
> need to read and understand the structure of jpeg deeply. Can anyone guide
> me where to look at ? 

I can give a brief intro, but if you need more information, you should
have a look at

@Book{Penn93j,
  author =	 {W. B. Pennebaker and J. L. Mitchell},
  title = 	 {JPEG Still Image Compression Standard},
  publisher = 	 {Van Nostrand Reinhold},
  address =      {New York},
  year = 	 1993
}

@Article{Wall91s,
  author = 	 {G. K. Wallace},
  title = 	 {The JPEG Still Picture Compression Standard},
  journal = 	 {Communications of the ACM},
  year = 	 1991,
  volume =	 34,
  number =	 4,
  pages =	 {30--44}
}

JPEG first converts the image to be compressed into the YCbCr color 
space and breaks up each color plane into 8x8 blocks of pixels. Then, 
all blocks are transformed with the Discrete Cosine Transform. 
In a quantization step all DCT coefficients are divided by some 
predefined quantization values and rounded to the nearest integer 
(according to a "quality factor", the quantization values
can be scaled by a constant). The purpose of this process is
to modulate the influence of the different spectral components on the image. 
In particular, the influence of the highest DCT coefficients is
reduced: They are likely to be  dominated by noise and are not expected 
to contribute significant details to the picture.  The resulting quantized
DCT coefficients are compressed using an entropy coder (e.g., Huffman or 
arithmetic coding). In the JPEG decoding step
all DCT coefficients are dequantized, i.e. multiplied with the quantization
values which had been used in the encoding step. Afterwards an inverse DCT is
performed to reconstruct the data. The restored picture will be close to (but not
identical with) the original one; but if the quantization values were
set properly, there should be no noticeable difference for a human observer.

> like if I want to modify the DCT coefficient ,I dont
> know  which bit in .JPG   contains the dct coefficien or what section
> contains it.
> 

You can try to use the portable JPEG library, available at
ftp://ftp.uu.net/graphics/jpeg/, there are also numerous mirrors.
If you look for code performing the DCT transform, you can visit my
homepage. I have an old implementation of a DCT-based steganography
system there; you can cut-and-paste the DCT/IDCT routines from there.

Regards,
Stefan.

-- 
----------------------------------------------------------
Stefan Katzenbeisser        skatzenbeisser@acm.org
                            private: katz@mail.com
http://stud3.tuwien.ac.at/~e9625414
----------------------------------------------------------

watermarking reflector, sent by majordomo 1.94.4
to unsubscribe send e-mail to watermarking-request with "unsubscribe" in the body