率-失真优化的渐进几何压缩

提出一种率-失真优化的渐进几何压缩方法. 根据三维模型细节信息分布的不均匀性, 该方法将细节信息分块并对各块独立编码, 然后以一定位率下重构网格几何失真最小为准则, 将各块的位流优化组装成最终的码流, 从而在渐进传输时使有限的网络带宽能优先分配给那些细节信息较为丰富的块. 实验结果表明, 与渐进几何压缩方法 (Progressive geometry compression, PGC) 相比, 在低位率时本文方法重构网格的峰值信噪比 (Peak signal-to-noise ratio, PSNR) 提高了约 2.25dB. 此外, 该方法也为实现三维网格感兴趣区域编码提供了新的方案.

Rate-distortion Optimized Progressive Geometry Compression

A novel scheme for rate-distortion(R-D)optimized progressive geometry compression is proposed.According to the non-uniform distribution of the model details,the model details are grouped into blocks,and each block is encoded independently.The encoded bit-streams of the blocks are assembled into a code stream based on R-D optimization,which ensures allocating bitrates to the blocks with richer details preferentially.Experimental results show that the proposed algorithm,compared with the well-known PGC,provides PSNR improvements by about 2.25 dB at low bitrate.Moreover, the scheme provides a new approach to realize region of interest(ROI)coding of 3D meshes efficiently.