Deep Part Induction from Articulated Object Pairs

Introduction

This work is based on our SIGGRAPH Asia 2018 paper. You can find the arXiv version of the paper here. In this repository, we release the training and evaluation code, data as well as the pre-trained models.

To Get Start

Download training and validation data from

https://shapenet.cs.stanford.edu/ericyi/data_partmob.zip

Compile the PointNet++ code in "pointnet2"

Train the corrspondence proposal and the flow module through

python train.py --stage 1

Train the hypothesis generation and the verification submodule through

python train.py --stage 2

Train the hypothesis selection submodule through

python train.py --stage 3

Evaluate the model through

python evaluation.py

You can also download the pretrained model from the following link:

https://shapenet.cs.stanford.edu/ericyi/pretrained_model_partmob.zip

License

Our code and data are released under MIT License (see LICENSE file for details).