Facta Universitatis, Series: Automatic Control and Robotics, Vol 13, No. 2 (2014)

A RUNLENGTH CODED ADAPTIVE QIM FOR THE CULTURAL HERITAGE 3D MODELS AUTHENTICATION

Dejan Rančić, Bata Vasić

Abstract


We describe a simple and effective quantization scheme for a digital 3D cultural heritage models authentication and watermarking. It is based on runlength coding which converts a class of deletion channels that have infinite memory into memoryless channels. We consider a novel application of this technique in multimedia watermarking and authentication using quantization index modulation operating on the three dimensional mesh vertices, which are invariant to geometric and topological transformation. Vector of the vertex indices is extracted from a huge digital 3D model using our powerful vertex extraction tool. The coding recovers the data hidden in the vertices removed by the process of mesh simplification.