【摘要】 对基于视频的点云压缩方案（V-PCC）编码树单元（CTU）级的码率控制中给未占用像素分配码率的问题，提出一种基于占用像素的几何图码率控制改进算法。对于不含未占用像素的CTU,根据CTU占用像素比例与码率分配权重间的相关性，设计了基于占用像素比例的码率分配权重方案，通过建立几何图码率、几何图失真、点云重建几何失真间的数学关系，求解CTU的码率分配权重；对于不含占用像素的CTU,不给其分配编码比特。对于码率分配算法的初始参数，依据点云序列重新统计更新。实验结果表明：几何图在全帧内配置下，与原始码率控制算法相比，所提算法的Geom.BD-GeomRate中D₁、D₂平均降低了18.5%和18.8%。更多还原

【Abstract】 A geometric graph rate control improvement algorithm based on occupied pixels is proposed to address the issue of assigning rate to unoccupied pixels in the Coding Tree Unit（CTU） level rate control of Video-based Point Cloud Compression（V-PCC） scheme. For CTUs without unoccupied pixels, a rate allocation weight scheme based on the proportion of occupied pixels and the correlation between rate allocation weights was designed. By establishing mathematical relationships between geometric image rate, geometric image distortion, and point cloud reconstruction geometric distortion, the rate allocation weights of CTUs were solved. For CTUs that do not contain occupied pixels, no encoding bits are assigned to them. For the initial parameters of the bit rate allocation algorithm, they are recalculated and updated based on the point cloud sequence. The experimental results show that, under full frame configuration, the proposed algorithm outperforms the original rate control algorithm in terms of D₁ and D₂ in Geom.BD-GeomRate decreased by an average of 18.5% and 18.8%, respectively.更多还原