| Budget Amount *help |
¥3,100,000 (Direct Cost: ¥3,100,000)
Fiscal Year 2006: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2005: ¥900,000 (Direct Cost: ¥900,000)
Fiscal Year 2004: ¥1,300,000 (Direct Cost: ¥1,300,000)
|
|---|
| Research Abstract |
Recently copyright issues for digital contents are becoming a serious problem. Especially when the copyrighted digital contents are exposed to the internet, they are an easy target for malicious parties to produce pirate digital contents for unauthorized sales. Digital fingerprinting, defined as a process to extract data called fingerprint from a digital content to protect the copyright of the owners, is becoming an active research topic.
A line of curvature is a curve on a surface that has tangent which are principal direction at all its points. The principal direction at a given point are those directions for which the normal curvature takes on minimum and maximum value, and those direction are orthogonal. However this orthogonality is lost in the point that is called umbilic because a maximum and minimum principal curvature become equal. Umbilic is classified into general umbilics and non-general umbilics. Then three generic features are called lemon, star, monster based on the patte
… More
rn of the net of lines of curvature at generic umbilics, we use those pattern and positions as digital fingerprinting.
We developed a method to judge the similarities between the original and the copied polygonal mesh models statistically based on the intrinsic surface properties of the polygonal meshes. In this study, to evaluate intrinsic properties of the polygonal meshes, we suggest using the subdivision surface technique. Subdivision surface is a method to represent a smooth shape in terms of a coarse polygonal model. As the number of subdivisions goes to infinity, the polygonal mesh converges to a smooth limit surface, and we extract intrinsic properties from the smooth limit surface.
The method consists of four phases:
(i) Matching polygonal models: We match the centroids, principal directions of the two models, and if necessary one of the model is uniformly scaled based on the relative volumes or surface areas of the two models.
(ii) Weak test: After matching, distances of the two models are compared at grid points.
(iii) Intermediate test: principal directions of curvature are compared at grid points.
(iv) Strong test: If umbilics exist, pattern and positions are compared These umbilics act as fingerprints on the surface and remain unchanged if the boundary surface undergoes small deformations.
Overall, this report provides computational methods to relate the copied model to the original model statistically, and shows the effectiveness of the method through examples. Less
|
