Endo-2DTAM: Advancing Dense Endoscopic Reconstruction with Gaussian Splatting-driven Surface Normal-aware Tracking and Mapping

Yiming Huang *, Beilei Cui *, Long Bai *, Zhen Chen, Jinlin Wu, Zhen Li, Hongbin Liu, Hongliang Ren

ICRA 2025

|| Paper || Arxiv ||

🛠️ Requirements

You can install them following the instructions below.

conda create -n endo2dtam python=3.10 # recommended
conda activate endo2dtam
# torch and cuda version according to your env and device
pip install torch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu118
pip install -r requirements.txt

git clone https://github.com/hbb1/diff-surfel-rasterization.git
pip install diff-surfel-rasterization

git clone https://gitlab.inria.fr/bkerbl/simple-knn.git
pip install simple-knn

Latest version is recommended for all the packages unless specified, but make sure that your CUDA version is compatible with your pytorch.

Hardware: Ubuntu20.04+RTX4090

⚓ Preparation

We use the C3VD dataset. You can use the scripts in data/prepeocess_c3vd to preprocess the dataset. We also provide the preprocessed dataset by EndoGSLAM: Google Drive.

After you get prepared, the data structure should be like this:

- data/
  |- C3VD/
    |- cecum_t1_b/
      |- color/
      |- depth/
      |- pose.txt
    |- cecum_t3_a/
- scripts/
  |- main.py
- utils/
- other_folders/
- readme.md

Training and Evaluation

bash slam.sh

bash eval.sh

Cite

@article{huang2025advancing,
  title={Advancing Dense Endoscopic Reconstruction with Gaussian Splatting-driven Surface Normal-aware Tracking and Mapping},
  author={Huang, Yiming and Cui, Beilei and Bai, Long and Chen, Zhen and Wu, Jinlin and Li, Zhen and Liu, Hongbin and Ren, Hongliang},
  journal={arXiv preprint arXiv:2501.19319},
  year={2025}
}

Acknowledgements

We would like to acknowledge the following inspiring work:

• 3DGS (Bernhard Kerbl et al.)
• 2DGS (Huang et al.)
• SplaTAM (Nikhil Keetha et al.)
• EndoGSLAM (Wang et al.)