Efficient coding for dynamic 3-D mesh with vertices structuring and connectivity-oriented multiresolution decomposition

In this paper, we propose a new coding method for dynamic 3-D polygonal meshes with invariant connectivity between frames. This coding method is based on a wavelet transform in time domain and a multiresolution decomposition in space domain. In the time domain, with a wavelet transform taking into account a correlation between frames, a temporal redundancy is reduced. In the space domain, using a structuring by acquiring surrounding vertices in order to apply the multiresolution decomposition, polygonal meshes that have a complex connectivity are structured in 2-D plane, and then, the structured 2-D data are compressed with a connectivity-oriented multiresolution decomposition and SFQ (space–frequency quantization). As a result of applying the proposed method and comparing it with the widely used coding scheme Dynapack, we confirmed a better coding efficiency especially in low bit rate.

► A new coding method for dynamic 3-D polygonal meshes with invariant connectivity between frames is proposed.
► The temporal redundancy is reduced with a wavelet transform.
► The structured 2-D data are encoded using a connectivity-oriented multiresolutional decomposition and SFQ.
► The proposed method gave a better coding performance especially in low bit rate comparing to Dynapack.