This is a MXNet implementation of PointCNN. It is as efficent as the origin Tensorflow implemetation and achieves same accuracy on both classification and segmentaion jobs. See the following references for more information:

"PointCNN"
Yangyan Li, Rui Bu, Mingchao Sun, Baoquan Chen
arXiv preprint arXiv:1801.07791, 2018.

https://arxiv.org/abs/1801.07791

We've tested code on MNIST only.

python ./download_datasets.py -d mnist -f ./
python ./prepare_mnist_data.py -f ./mnist/zips
python ./pointcnn_cls.py

Our code is released under MIT License (see LICENSE file for details).