|
|
Watermarking 3D Models
Honzátková, Tereza ; Herout, Adam (referee) ; Španěl, Michal (advisor)
Goals of this work are to summarize the process and existing methods of 3D models watermarking, to choose relevant methods for embedding/extraction of a watermark to/from 3D models. The final solution is based on Spectral Mesh Compression technique. This method is robust, imperceptible and informed. The embedded watermark is detectable and robust against geometrical transformations, added random noise, mesh smoothing and simplification, compression and cropping. The realized tool allows a user to embedding watermark into the mesh, modify the mesh and extraction of a watermark from mesh. Testing was conducted on a set of 5 models differing in size, shape and density of points. The resulting tool inserts a watermark composed of four characters that can detect both undamaged models and the models affected. Among the attacks against which the watermark is robust, include transformations, rotation, scale, adaptive random noise, and combination of this.
|
|
|
Watermarking 3D Models
Honzátková, Tereza ; Herout, Adam (referee) ; Španěl, Michal (advisor)
Goals of this work are to summarize the process and existing methods of 3D models watermarking, to choose relevant methods for embedding/extraction of a watermark to/from 3D models. The final solution is based on Spectral Mesh Compression technique. This method is robust, imperceptible and informed. The embedded watermark is detectable and robust against geometrical transformations, added random noise, mesh smoothing and simplification, compression and cropping. The realized tool allows a user to embedding watermark into the mesh, modify the mesh and extraction of a watermark from mesh. Testing was conducted on a set of 5 models differing in size, shape and density of points. The resulting tool inserts a watermark composed of four characters that can detect both undamaged models and the models affected. Among the attacks against which the watermark is robust, include transformations, rotation, scale, adaptive random noise, and combination of this.
|
guest ::
