Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Learning Fuzzy Set Representations of Partial Shapes on Dual Embedding Spaces

Minhyuk Sung, Anastasia Dubrovina, Vladimir G. Kim, and Leonidas Guibas
SGP 2018 (Symposium on Geometry Processing)
[Project] | [arXiv]



  author = {Sung, Minhyuk and Dubrovina, Anastasia, and Kim, Vladimir G. and Guibas, Leonidas},
  title = {Learning Fuzzy Set Representations of Partial Shapes on Dual Embedding Spaces},
  Journal = {Computer Graphics Forum (Proc. of Symposium on Geometry Processing (SGP))}, 
  year = {2018}


This neural-network-based framework analyzes an uncurated collection of 3D models from the same category and learns two important types of semantic relations among full and partial shapes: complementarity and interchangeability. The former helps to identify which two partial shapes make a complete plausible object, and the latter indicates that interchanging two partial shapes from different objects preserves the object plausibility. These two relations are modeled as fuzzy set operations performed across the dual partial shape embedding spaces, and within each space, respectively, and jointly learned by encoding partial shapes as fuzzy sets in the dual spaces.


  • Python-gflags (tested with ver. 3.1.2)
  • Networkx (tested with ver. 2.1)
  • Numpy (tested with ver. 1.14.2)
  • Pandas (tested with ver. 0.23.0)
  • Scipy (tested with ver. 1.0.1)
  • TensorFlow-gpu (tested with ver. 1.4.0)

Reproducing paper results

Data download/generation

Download ComplementMe component point cloud data:

cd data
cd ..

The partial shapes used in the paper can be generated from the components by the following script:

cd data
cd ..

Pretrained model download

We provide pretrained models for all categories. Run:

cd fuzzy_set_dual/experiments
cd ../../

Network training

You can also easily train the network from scratch. Specify a category of shapes to train (set one of directory names in data/components. e.g. Chair. Case-sensitive):

export synset=Chair

Move to the experiment directory:

cd fuzzy_set_dual/experiments

For learning complementarity, run:

./ --relative --train

For learning interchangeability, run:

./ --train

The trained models are stored in fuzzy_set_dual/experiments/($synset)/vanilla_100_centerize_relative and fuzzy_set_dual/experiments/($synset)/vanilla_100_centerize directories, respectively.


The retrieval of complementary and interchangeable partial shapes are performed by runnning the same script without the --train option.

For retrieving with all test partial shapes in all categories as quries, run:

Regenerating paper figures/tables.

We provide script files in figures regenerating results in the paper figures and tables.

The script files require ComplementMe mesh data, and the mesh data is provided by the authors upon request (see here). Download the mesh data in data directory and unzip them.

For running the script files, first download/generate all data, download pretrained models, and run the batch retrieval scripts as described above. The outputs of each figure script are stored as mesh files, and the results of compared methods are not generated. Also, the position of retrieved complement partial shapes are not predicted since it is not a part of this work.

Note that the scripts have a dependency with MeshLab. Ubuntu users can install with apt-get:

sudo apt-get install meshlab


This project is developed based on the code in ComplementMe. The files in network_utils (except are directly brought from the PointNet++.


This code is released under the MIT License. Refer to LICENSE for details.


No description, website, or topics provided.







No releases published


No packages published