CP-Net: Contour-Perturbed Reconstruction Network for Self-Supervised Point Cloud Learning
Usage
Requirement
- Python 3
- Pytorch 0.4
- CMake > 2.8
Note: The code is not not compatible with Pytorch >= 1.0 due to the C++/CUDA extensions.
Building C++/CUDA Extensions for PointNet++
mkdir build && cd build
cmake .. && make
Training & Evaluation
Self-supervised pretraining:
bash self-supervised_pretrain.sh RSCNN ShapeNetPart
After pretraining, you can get the pretrained features with 'train_features_saved.h5' and 'test_features_saved.h5', which is the input of nect stage: semi-supervised part segmentation:
python semi-supervised-finetuning.py --sample_rate 0.05 --exp exp_name