Lossy Video Coding of V-DMC Displacements

This paper presents the compression of mesh displacement data based on lossy video coding that has been recently adopted into the emerging ISO/IEC 23090-29 Video-based dynamic mesh coding (V-DMC) standard. As its name indicates, the emerging V-DMC standard relies on video coding for the compression of dynamic meshes, specifies metadata and the decoding processes. More specifically, video coding is used in V-DMC to code texture (attribute) data and displacement (geometry) data. V-DMC has been designed on the principles of subdivision surfaces and the lifting scheme; the displacement data consists in residual information that is hierarchically predicted through Levels-of-Details (LoD). Due to this nature of the data, previous versions of V-DMC relied on lossless video coding of displacement data after quantization of the raw displacement signal, with the intuition that this leads to a better control of the quantization error. We show in this paper that signaling a linear scale factor while using a video codec in lossy mode provides better coding gains and other benefits to V-DMC.