|
Activity Number:
|
103
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #300890 |
|
Title:
|
Probabilistic Watermark Detection in Movies
|
|
Author(s):
|
Sam Behseta+ and Charles Lam and Robert Webb*+
|
|
Companies:
|
California State University, Bakersfield and California State University, Bakersfield and California Polytechnic State University
|
|
Address:
|
Department of Mathematics, Bakersfield, CA, 93311, , San Luis Obispo, CA, 93407 ,
|
|
Keywords:
|
Image Processing ; Nonparametric Bootstrap ; Watermarking ; Discrete Cosine Transforms
|
|
Abstract:
|
In this work, we propose different algorithms to approximate an original watermarked document, primarily a movie represented with a sequence of matrices. We consider the case of a collusion attack when a number of watermarked copies are available. We use the Bootstrap to construct point-wise confidence intervals for each matrix, represented via their discrete cosine transforms. We demonstrate, by way of an extensive simulation that in addition to stellar probabilistic coverage, Bootstrap machinery is highly efficient when the number of Bootstrap iterations, the sample size and the number of watermarked copies are of interest. Most importantly, the results suggest that the precision of our probabilistic methodology increases quickly when the number of watermarked copies are increased.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2008 program |
