Cascaded quantization based progressive 3D mesh compression

Abstract

We proposed an efficient progressive 3D mesh compression method supporting flexible quality scalability. In the proposed method, the mesh geometry prediction residuals are partitioned into a number of iterative layers. Each iterative layer is split into several quality layers using cascaded quantization and then encoded by context adaptive arithmetic codec (CABAC). All the quality layers are encoded and transmitted independently to enable better error resilience. To achieve better rate-distortion performance, the quantization parameter of the first quality layer is determined by the importance of the corresponding iterative layer. Simulation results demonstrate that the proposed method is able to provide better compression performance compared to the state-of-the-art coders.