Paper

Overview of our method.

This repository contains the implementation of the paper:

If you find our code or paper useful, please consider citing:

@inproceedings{li2024GridFormer,
      title={GridFormer: Point-Grid Transformer for Surface Reconstruction},
      author={Li, Shengtao and Gao, Ge and Liu, Yudong and Liu, Yu-Shen and Gu, Ming},
      booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
      year={2024}
    }

You can create an anaconda environment called gridformer using

conda env create -f environment.yaml
conda activate gridformer
pip install -r requirements.txt

Next, compile the extension modules. You can do this via

python setup.py build_ext --inplace

You can download the dataset (73.4 GB) by running the script from Occupancy Networks. After, you should have the dataset in data/ShapeNet folder.

You can download the preprocessed data (144 GB) using

bash scripts/download_data.sh

This script should download and unpack the data automatically into the data/synthetic_room_dataset folder.

Download ScanNet v2 data from the official ScanNet website. Then, you can preprocess data with: scripts/dataset_scannet/build_dataset.py and put into data/ScanNet folder.

We provide the boundary sampling python script for ShapeNet. You can run it with the preprocessed dataset used in ConvONet by:

python scripts/boundary_sampling.py

For other datasets or your own data, just change the path in the script.

Stage 1:

python train.py CONFIG.yaml

Stage 2:

python refine.py CONFIG.yaml

For available training options, please take a look at configs/default.yaml.

To generate meshes using a trained model, use

python generate.py CONFIG.yaml

where you replace CONFIG.yaml with the correct config file.

For evaluation of the models, we provide the script eval_meshes.py. You can run it using:

python eval_meshes.py CONFIG.yaml

The script takes the meshes generated in the previous step and evaluates them using a standardized protocol. The output will be written to .pkl/.csv files in the corresponding generation folder which can be processed using pandas.

We provide the pre-trained models for ShapeNet 3k with noise and Synthetic Rooms 10k with noise. You can download the models from the link.

The code is largely based on ConvONet and ALTO. Many thanks to the authors for opensourcing the codebase.