Topological approaches to 3D mesh watermarking
The recent rapid growth of digital media contents and the increase use of online services have triggered the need for multimedia protection. Watermarking plays an important role to solve the problem of unauthorized replication. Watermarking can be defined as the process of embedding data called &quo...
Main Author: | |
---|---|
Format: | Others |
Published: |
2008
|
Online Access: | http://spectrum.library.concordia.ca/975863/1/MR40902.pdf Abbas, Abdullah Omer A <http://spectrum.library.concordia.ca/view/creators/Abbas=3AAbdullah_Omer_A=3A=3A.html> (2008) Topological approaches to 3D mesh watermarking. Masters thesis, Concordia University. |
Summary: | The recent rapid growth of digital media contents and the increase use of online services have triggered the need for multimedia protection. Watermarking plays an important role to solve the problem of unauthorized replication. Watermarking can be defined as the process of embedding data called "watermark" into a digital object without making changes to the quality of the host substantially. The digital object could be an image, video, or audio. The watermark is used as a signature to prove ownership and can only be detected or extracted by the owner. This thesis is devoted to three robust watermarking techniques that we have developed for copyright protection. The first watermarking technique inserts a binary number into a set of critical points of a digital elevation map (DEM). Our method starts by extracting the critical points from a DEM depending on the important topographic features of the terrain. Then we embed the fingerprints into the coordinate values of all the critical points. The second watermarking technique partitions a 3D model into sub-meshes, then apply the eigen-decomposition to the Laplace-Beltrami matrix of each sub-mesh, followed by computing the hash value of each sub-mesh. The hash value is defined in terms of the entropy of each submesh. The last watermarking technique applies content-based hashing algorithm to the DEM using higher-order statistics and discrete wavelet transform to define the image fingerprint |
---|