DeepJoin

Code for "DeepJoin: Learning a Joint Occupancy, Signed Distance, and Normal Field Function for Shape Repair."
[arXiv] Published at SIGGRAPH Asia 2022.

Input	Output

If you find our work helpful please cite us:

@article{lamb2022deepjoin,
  title={DeepJoin: Learning a Joint Occupancy, Signed Distance, and Normal Field Function for Shape Repair},
  author={Lamb, Nikolas and Banerjee, Sean and Banerjee, Natasha Kholgade},
  journal={ACM Transactions on Graphics (TOG)},
  volume={41},
  number={6},
  pages={1--10},
  year={2022},
  publisher={ACM New York, NY, USA}
}

Installation

Code tested using Ubutnu 18.04 and python 3.8.0. Note that you need to have the following apt dependencies installed.

sudo apt install python3.8-distutils python3.8-dev libgl1 libglew-dev freeglut3-dev

Clone the repo.

https://github.com/Terascale-All-sensing-Research-Studio/DeepJoin.git
cd DeepJoin

We recommend using virtualenv. The following snippet will create a new virtual environment, activate it, and install deps.

sudo apt-get install virtualenv && \
virtualenv -p python3.8 env && \
source env/bin/activate && \
pip install -r requirements.txt && \
./install.sh && \
source setup.sh

Issues with compiling pyrender are typically solved by upgrading cython: pip install --upgrade cython.

If you want to run the fracturing and sampling code, you'll need to install pymesh dependencies:

./install_pymesh.sh

Quickstart Inference

If you just want to try out inference, run the following script with the example file. This will infer a restoration and create a gif.

cd deepjoin
./scripts/infer_quick.sh experiments/mugs/specs.json ../example_files/fractured_mug.obj

You should get a gif that looks like this one!

Data Preparation

See fracturing/README.md.

Training

Navigate into the deepjoin directory.

cd deepjoin

Each experiment needs a corresponding directory with a "specs.json" file. You can find an example at deepjoin/experiments/mugs.

To train, run the training python script with the path to an experiment directory.

python python/train.py -e experiments/mugs

Inference

Navigate into the deepjoin directory.

cd deepjoin

Inference (and related operations) is done in four steps:

1. Infer latent codes.
2. Reconstruct meshes.
3. Generate renders.
4. Evaluate meshes.

Each experiment needs a corresponding directory with a "specs.json" file. You can find an example at deepjoin/experiments/mugs.

To infer:

./scripts/infer.sh experiments/mugs

Data is saved in the experiment directory passed to the reconstruction script, under a Reconstructions subdirectory. For example, results for the mugs example will be stored in deepjoin/experiments/mugs/Reconstructions/ours/. Meshes are stored in the Meshes subdirectory. A render of all the results is stored in the top-level reconstruction directory.