EE368 Project proposal

Name : Kyoungsik Yu


Title: Efficient attack filter for Digital watermarking

  In this project, I'll investigate the efficient attack and counter-attack 
method for imperceptible digital watermarking technique concentrating on 
decision feedback filtering. Digital watermarking is a promising candidate 
for protecting the copyright of digital multimedia contents.

  So far, most of conventional attack methods were assuming an IIR 
(infinite-lenght impulse response) linear filtering to implement the attack 
filter. However, this is hard to realize in DSP board and real-time processing.
 Especially, IIR approach seems to be vulnearable to the quantization effect 
of data.
  FIR (finite-length impulse response) filter attack method can be faster than
 existing attack scheme with almost same performance, allowing an efficient
 filtering and a real-time computation for a DSP implementation.

  To prove this, one-dimensional image and watermarked sequences will be used 
for simulation and analysis for simplicity. (It can be extended to 
two-dimensional image data if time is allowed.) The effect of
 filter-coefficent quantization and filter-length will be analysed 
theoretically with simulation results.
 In addition, I'll explore non-linear filtering technique to remove the 
watermarking using decision feedback filter. Decision-feedback equalizer 
is a mature technique to equalize the passband spectrum in digital 
communication area. Theoretically, decision-feedback attack method can have 
a potential to have a better performance than the existing linear-filter 
attack. Dynamic programming to combat harzadous error propagation will be 
also implemented to get full advantages of decision feedback structure.



References
[1] J.K.Su and B. Girod, "On the robustness and impercemtibility of digital 
fingerprints", 1999


 +------------------------------------------+
 |            < Kyoungsik Yu >              |
 +------------------------------------------+
 | E-mail  : ksyu@leland.stanford.edu       |
 | Phone   :  [Home]  (650) 497-3571        |
 |           [Office] (650) 723-6104        |
 | Address :  [Home] 141D Escondido Village |
 |                    Stanford, CA, 94305   |
 |           [Office] 41C E.L.Ginzton Lab.  |
 |                    Stanford, CA, 94305   |
 +------------------------------------------+