A Low-Complexity Patch Segmentation in the V-PCC Encoder

Abstract

The video-based point cloud compression(V-PCC) is the compression standard of the dynamic dense point cloud established by ISO/IEC MPEG-I in 2021. It is designed to achieve significant data size reduction for the dynamic dense point cloud. However, the patch generation process in the V-PCC encoding is computationally very complex. To solve this problem, we propose a fast algorithm for patch partitioning using a look-up table instead of the k-d tree search. Experimental results show the reduction of the computational complexity of the patch segmentation process by over 40% while achieving the same encoding results as the original method.