Skip to content
3D Point Capsule Networks
Branch: master
Clone or download
tolgabirdal and yongheng1991 first test commit: changed the citation to the CVPR one.
first test commit: changed the citation to the CVPR one.

fixing the README.

fixing the README.

fixing cont.

fixing cont.

Update README.md

removed the dash between point and capsule

tested with the pre-trained model

add interpolation

rm readme tmp

Delete README.md

add readme back

Update README.md

modify the readme
Latest commit 1304d4b Apr 3, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
apps first test commit: changed the citation to the CVPR one. Apr 12, 2019
dataloaders initial commit Apr 2, 2019
dataset initial commit Apr 2, 2019
docs initial commit Apr 2, 2019
mini_example/AE first test commit: changed the citation to the CVPR one. Apr 12, 2019
models
README.md first test commit: changed the citation to the CVPR one. Apr 12, 2019

README.md

3D Point Capsule Networks

Created by Yongheng Zhao, Tolga Birdal, Haowen Deng, Federico Tombari from TUM.

This repository contains the implementation of our CVPR 2019 paper 3D Point Capsule Networks. In particular, we release code for training and testing a 3D-PointCapsNet network for classification, reconstruction, part interpolation and extraction of 3d local descriptors as well as the pre-trained models for quickly replicating our results.

Part Interpolation

Distribution of capsule reconstruction

Abstract

In this paper, we propose 3D point capsule networks, an auto-encoder designed to process sparse 3D point clouds while preserving spatial arrangements of the input data. 3D capsule networks arise as a direct consequence of our novel unified 3D auto-encoder formulation. Their dynamic routing scheme and the peculiar 2D latent space deployed by our approach bring in improvements for several common point cloud-related tasks, such as object classification, object reconstruction and part segmentation as substantiated by our extensive evaluations. Moreover, it enables new applications such as part interpolation and replacement.

Citation

If you find our work useful in your research, please consider citing:

	  @inproceedings{zhao20193d, 
		author={Zhao, Yongheng and Birdal, Tolga and Deng, Haowen and Tombari, Federico}, 
		booktitle={Conference on Computer Vision and Pattern Recognition (CVPR)}, 
		title={3D Point 
		Capsule Networks}, 
		organizer={IEEE/CVF},
		year={2019}
	  }

Installation

The code is based on PyTorch. It has been tested with Python 3.6, PyTorch 1.0.0, CUDA 9.2(or higher) on Ubuntu 16.04. (You can also use PyTorch 0.4.0 but you need to replace the Chamfer distance package with the original nndistance.)

Install h5py for Python:

  sudo apt-get install libhdf5-dev
  sudo pip install h5py

Install Chamfer distance package: (Be aware of the PyTorch 1.0.1. It may have a problem for building this cuda package.)

  cd models/nndistance
  python build.py install

To visualize the training process in PyTorch, consider installing TensorBoard.

To visualize the reconstructed point cloud, consider installing Open3D.

Datasets

ShapeNetPart Dataset

  cd dataset
  bash download_shapenet_part16_catagories.sh

ShapeNet Core with 13 categories (refered from AtlasNet.)

  cd dataset
  bash download_shapenet_core13_catagories.sh

ShapeNet Core with 55 categories (refered from FoldingNet.)

  cd dataset
  bash download_shapenet_core55_catagories.sh

Pre-trained model

You can download the pre-trained models here.

Usage

A Minimal Example

We provide an example demonstrating the basic usage in the folder 'mini_example'.

To visualize the reconstruction from latent capsules with our pre-trained model:

  cd mini_example/AE
  python viz_reconstruction.py --model ../../checkpoints/shapenet_part_dataset_ae_200.pth

To train a point capsule auto encoder with ShapeNetPart dataset by yourself:

  cd mini_example/AE
  python train_ae.py

Point Capsule Auto Encoder

To train a point capsule auto encoder with another dataset:

  cd apps/AE
  python train_ae.py --dataset < shapenet_part, shapenet_core13, shapenet_core55 >

To monitor the training process, use TensorBoard by specifying the log directory:

  tensorboard --logdir log

To test the reconstruction accuracy:

  python test_ae.py  --dataset < >  --model < >
e.g. 
  python test_ae.py --dataset shapenet_core13 --model ../../checkpoints/shapenet_core13_dataset_ae_230.pth

To visualize the reconstructed points:

  python viz_reconstruction.py --dataset < >  --model < >
e.g. 
  python viz_reconstruction.py --dataset shapenet_core13 --model ../../checkpoints/shapenet_core13_dataset_ae_230.pth

Transfer Learning and Semi Supervised Classification

To generate latent capsules from a pre-trained model and save them into a file:

  cd apps/trasfer_cls
  python save_output_latent_caps_in_file.py --dataset < >  --model < >  --save_training  # process and save training dataset
  python save_output_latent_caps_in_file.py --dataset < >  --model < >   # process and save testing dataset

To train and test the liner SVM classifier with the pre-trained AE model and pre-saved latent capsules:

  python train_and_test_svm_cls_from_pre-saved_latent_caps.py --dataset < >  --model < >

The AE model and latent capsules are obtained from different datasets in order to test the performance of classification under transfer.

Training a Liner SVM classifier with a limited part of the training data and testing with the complete test data:

  python train_and_test_svm_cls_from_pre-saved_latent_caps.py --dataset < >  --model < > --percent_training_dataset < 5, 10 ...>
e.g.
  python train_and_test_svm_cls_from_pre-saved_latent_caps.py --dataset shapenet_part  --model ../../checkpoints/shapenet_part_dataset_ae_200.pth --percent_training_dataset 10

Part Segmentation

To generate latent capsules with the part label from a pre-trained model and save them into a file (The model is also trained with shapenet-part dataset):

  cd apps/part_seg
  python save_output_latent_caps_with_part_label.py --dataset shapenet_part  --model < >  --save_training  # process and save training dataset
  python save_output_latent_caps_with_part_label.py --dataset shapenet_part  --model < >   # process and save testing dataset

To train a capsule-wise part segmentation with a specific amount of training data:

  python train_seg.py --model < > --percent_training_dataset < 5, 10 ...>
e.g.
  python train_seg.py --model ../../checkpoints/shapenet_part_dataset_ae_200.pth --percent_training_dataset 1

To evaluate and visualize the part segmentation: (--model < pre-trained model of point capsule auto encoder >; --part_model )

 python eva_seg.py --model < >  --part_model < > --class_choice < >
e.g.
 python eva_seg.py --model ../../checkpoints/shapenet_part_dataset_ae_200.pth  --part_model ../../checkpoints/part_seg_1percent.pth  --class_choice Airplane

Part Interpolation and Replacement

To visualize the part interpolation in open3D:

 python part_interplation.py --model < >  --part_model < > --class_choice < >
e.g.
 python part_interplation.py --model ../../checkpoints/shapenet_part_dataset_ae_200.pth  --part_model ../../checkpoints/part_seg_100percent.pth  --class_choice Airplane

To visualize the part replacement in open3D:

 python part_replacement.py --model < >  --part_model < > --class_choice < >

3D Local Feature Extraction

to be continued...

License

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

Reference

The chamfer distance package is based on nndistance. The necessary modifications have been done to this repository in order to run it with PyTorch 1.0.0.

The capsule layer is based upon and modified from Capsule-Network-Tutorial

Our capsule decoder is based upon the decoder of AtlasNet.

You can’t perform that action at this time.