Improved grid refine segmentation for 3D point cloud in video-based point cloud compression (V-PCC)

For an immersive visual communication experience, it is essential to enable technologies that can capture and transmit point clouds capable of accurately depicting a photorealistic impression. The MPEG standardization group has devised a method for compressing video-based point clouds called Video-based Point Cloud Compression (V-PCC). In the V-PCC process, it first projects a 3D point cloud onto 2D planes called 3D patch generation, consisting of normal estimation, initial segmentation, grid refine segmentation (GRS), and patch segmentation. Afterwards, the High-Efficiency Video Coding (HEVC) standard is utilized for 2D video compression. However, the GRS is the most time-consuming process in 3D patch generation; a state-of-the-art method has evolved to address this issue. In this paper, we propose an improved approach that eliminates redundant execution steps and speeds up the optimization process of GRS by predicting the convergence of the determined point cloud projection plane during iterations. Our experimental results show that our approach is more persuasive, reducing execution time by up to 15% with slight increases or decreases in the BD rate compared to the state-of-the-art method.