🌟 MIND: Material Interface Generation from UDFs for Non-Manifold Surface Reconstruction

📖 Overview

This repository hosts the official code and resources for the paper "MIND: Material Interface Generation from UDFs for Non-Manifold Surface Reconstruction" published at NIPS 2025.

We propose a novel algorithm to extract non-manifold mesh from unsigned distance fields.

⚙️ Requirements and Installation

This project requires Python [Version] or higher.

Prerequisites

• We need pytorch installed.

Installation

1. Clone the repository:

   git clone https://github.com/jjjkkyz/MIND.git
   cd MIND
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Install:

   pip install -e .
   

🚀 Usage

    from mldf import MIND


    # some neural network
    udf_func = your_input
    resolution = 256

    # for more paremeters, see the source code
    mind_mesh_extractor = MIND(udf_func, resolution)

    mesh = mind_mesh_extractor.run()

📝 Citation

If you find this repository or our paper useful, please cite our work:

    @inproceedings{
        chen2025mind,
        title={{MIND}: Material Interface Generation from {UDF}s for Non-Manifold Surface Reconstruction},
        author={Xuhui Chen and Fei Hou and Wencheng Wang and Hong Qin and Ying He},
        booktitle={The Thirty-ninth Annual Conference on Neural Information Processing Systems},
        year={2025},
        url={https://openreview.net/forum?id=4lR9OhAisI}
    }

---

📄 License

This project is licensed under the MIT License