GRNet: Geometry Restoration for G-PCC Compressed Point Clouds Using Auxiliary Density Signaling

The lossy Geometry-based Point Cloud Compression (G-PCC) inevitably impairs the geometry information of point clouds, which deteriorates the quality of experience (QoE) in reconstruction and/or misleads decisions in tasks such as classification. To tackle it, this work proposes GRNet for the geometry restoration of G-PCC compressed large-scale point clouds.

• 2023.11.27 Our paper has been accepted by TVCG. [paper]

Requirments

• python3.7 or 3.8
• cuda10.2 or 11.0
• pytorch1.7 or 1.8
• MinkowskiEngine 0.5 or higher (for sparse convolution)
• tmc3 v21 (for G-PCC compression) https://github.com/MPEGGroup/mpeg-pcc-tmc13

We recommend you to follow https://github.com/NVIDIA/MinkowskiEngine to setup the environment for sparse convolution.

Training

TODO

Testing

chmod a+x ./tmc3

python test_solid.py --ckpts='ckpts_path' --GT_dir='GT_path' --last_kernel_size= --resolution= --posQuantscale= 

python test_dense.py --ckpts='ckpts_path' --GT_dir='GT_path' --last_kernel_size= --resolution= --posQuantscale= 

python test_dense_offset.py --ckpts='ckpts_path' --GT_dir='GT_path' --resolution= --posQuantscale= 

python test_sparse.py --ckpts='ckpts_path' --GT_dir='GT_path' --resolution= --posQuantscale=