Seyed Saber Mohammadi, Yiming Wang, Alessio Del Bue. PointView-GCN: 3D shape classification with multi-view point clouds. You can find IEEE version of the paper here.
If you find our work useful in your research, please consider citing:
@inproceedings{mohammadi2021pointview,
title={Pointview-GCN: 3D Shape Classification With Multi-View Point Clouds},
author={Mohammadi, Seyed Saber and Wang, Yiming and Del Bue, Alessio},
booktitle={2021 IEEE International Conference on Image Processing (ICIP)},
pages={3103--3107},
year={2021},
organization={IEEE}
}
You can find our dataset with partial single-view PCDs generated from benchmark dataset ModelNet40. Plese download the dataset, creat a directory named "single_view_modelnet" and put it under "data" directory.
First use the pre-trained model to extract the features from each single-view PCD:
cd Feature_extraction
python main.py
Then apply the GCN to aggregate and classify the features:
cd GCN
python main.py
You can also train the backbone from scratch:
cd PointNet++
python main.py
First download the normalize version of ModelNet40 dataset ModelNet40_normalized and put it under the "data" directory. then run the following comment:
cd dataset_rendering
python dataset_capturing.py --out-split-dir /train/ && python dataset_capturing.py --out-split-dir /test/
Note that, since the dataset generation takes a huge amount of the time, we provided the final version of the generated single-view PCDs.