Created by Xingyu Liu, Mengyuan Yan and Jeannette Bohg from Stanford University.
If you find this work useful in your research, please cite:
@inproceedings{liu2019meteornet,
title={MeteorNet: Deep Learning on Dynamic 3D Point Cloud Sequences},
author={Xingyu Liu and Mengyuan Yan and Jeannette Bohg},
booktitle={ICCV},
year={2019}
}Understanding dynamic 3D environment is crucial for robotic agents and many other applications. We propose a novel neural network architecture called MeteorNet for learning representations for dynamic 3D point cloud sequences. Different from previous work that adopts a grid-based representation and applies 3D or 4D convolutions, our network directly processes point clouds. We propose two ways to construct spatiotemporal neighborhoods for each point in the point cloud sequence. Information from these neighborhoods is aggregated to learn features per point. We benchmark our network on a variety of 3D recognition tasks including action recognition, semantic segmentation and scene flow estimation. MeteorNet shows stronger performance than previous grid-based methods while achieving state-of-the-art performance on Synthia. MeteorNet also outperforms previous baseline methods that are able to process at most two consecutive point clouds. To the best of our knowledge, this is the first work on deep learning for dynamic raw point cloud sequences.
Install TensorFlow. The code is tested under TF1.9.0 GPU version, g++ 5.4.0, CUDA 9.0 and Python 3.5 on Ubuntu 16.04. There are also some dependencies for a few Python libraries for data processing and visualizations like cv2. It's highly recommended that you have access to GPUs.
The TF operators are included under tf_ops, you need to compile them first by make under each ops subfolder (check Makefile) or directly use sh command_make.sh. Update arch in the Makefiles for different CUDA Compute Capability that suits your GPU if necessary.
The code for action recognition experiments on MSRAction3D dataset is in action_cls/. Please refer to action_cls/README.md for more information on data preprocessing and experiments.
The code for semantic segmentation experiments on Synthia dataset is in semantic_seg/. Please refer to semantic_seg/README.md for more information on data preprocessing and experiments.
Note that only direct grouping models are released for now. Chain-flowed models will be released soon.
To be released. Stay tuned!
To be released. Stay tuned!
Our code is released under MIT License (see LICENSE file for details).