This is the official repo of the ICML 2024 paper "Toward Availability Attacks in 3D Point Clouds".
Create environments for FC-EM:
conda create -n fc-em python=3.9
pip install -r requirements.txt
conda activate fc-emhttps://shapenet.cs.stanford.edu/media/modelnet40_ply_hdf5_2048.zip
https://hkust-vgd.ust.hk/scanobjectnn/
https://drive.google.com/drive/folders/1yjLdofRRqyklgwFOC0K4r7ee1LPKstPh/IntrA.zip
https://faces.dmi.unibas.ch/bfm/bfm2017.html
After download and uncompress dataset, move them to /data folder.
python main.py --exp-config configs/modelnet40/fc_errmin/fc_em_gen.yaml --gpu-id {gpu} --chamf_coeff {beta}
python main.py --exp-config configs/modelnet40/fc_errmin/fc_em_eval.yaml --gpu-id {gpu} --chamf_coeff {beta} python main.py --exp-config configs/extra/scan/scan_gen.yaml --gpu-id {gpu} --chamf_coeff {beta}
python main.py --exp-config configs/extra/scan/scan_eval.yaml --gpu-id {gpu} --chamf_coeff {beta} python main_medi.py --exp-config configs/extra/intra/intra_gen.yaml --gpu-id {gpu} --chamf_coeff {beta}
python main_medi.py --exp-config configs/extra/intra/intra_eval.yaml --gpu-id {gpu} --chamf_coeff {beta} python main_medi.py --exp-config configs/extra/face/bsm2017_gen.yaml --gpu-id {gpu} --chamf_coeff {beta}
python main_medi.py --exp-config configs/extra/face/bsm2017_eval.yaml --gpu-id {gpu} --chamf_coeff {beta} cd configs/defense
sh run_fc_em_defense.sh