The KITTI sequences are compressed in the G-PCC reference software TMC13v22 with predictive tree geometry coding. A simple and effective post-processing framework is proposed to enhance the geometry quality, which is also extended to LiDAR upsampling and denoising.
@article{liu2026post,
title={Post-Processing Geometry Enhancement for G-PCC Compressed LiDAR via Cylindrical Densification},
author={Liu, Wang and Li, Zhuangzi and Li, Ge and Ma, Siwei and Kwong, Sam and Gao, Wei},
journal={IEEE Transactions on Image Processing},
volume={35},
pages={1066--1081},
year={2026},
publisher={IEEE}
}