Skip to content
Associatively Segmenting Instances and Semantics in Point Clouds
Branch: master
Clone or download
WXinlong
Latest commit 327b5ff Mar 4, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data Initial commit Feb 26, 2019
meta Initial commit Feb 26, 2019
misc edit README Feb 26, 2019
models/ASIS add estimate_mean_ins_size.py file Mar 4, 2019
tf_ops Initial commit Feb 26, 2019
utils Initial commit Feb 26, 2019
LICENSE Initial commit Feb 26, 2019
README.md
collect_indoor3d_data.py Initial commit Feb 26, 2019
gen_h5.py
indoor3d_util.py Initial commit Feb 26, 2019

README.md

Associatively Segmenting Instances and Semantics in Point Clouds

[arXiv]

Overview

Dependencies

The code has been tested with Python 2.7 on Ubuntu 14.04.

Data and Model

  • Download 3D indoor parsing dataset (S3DIS Dataset). Version 1.2 of the dataset is used in this work.
python collect_indoor3d_data.py
python gen_h5.py
cd data && python generate_input_list.py
cd ..
  • (optional) Prepared HDF5 data for training is available here.

  • (optional) Trained model can be downloaded from here.

Usage

  • Compile TF Operators

    Refer to PointNet++

  • Training

cd models/ASIS/
ln -s ../../data .
sh +x train.sh 5
  • Evaluation
python eval_iou_accuracy.py

Note: We test on Area5 and train on the rest folds in default. 6 fold CV can be conducted in a similar way.

Citation

If our work is useful for your research, please consider citing:

@inproceedings{wang2019asis,
	title={Associatively Segmenting Instances and Semantics in Point Clouds},
	author={Wang, Xinlong and Liu, Shu and Shen, Xiaoyong and Shen, Chunhua, and Jia, Jiaya},
	booktitle={CVPR},
	year={2019}
}

Acknowledgemets

This code largely benefits from following repositories: PointNet++, SGPN, DGCNN and DiscLoss-tf

You can’t perform that action at this time.