Installation

sudo apt-get install libopenexr-dev
    
conda env create -f environment.yaml
conda activate enerf-slam

Run

Synthetic Dataset

The data can be generated by simulator. Then organize the data as follows.

Datasets/synthetic01
|---results
|      |----depth000000.png
|      |----depth000001.png
|      ...
|      ...
|      |----frame000000.png
|      |----frame000001.png
|      ...
|      ...
|---traj.txt

Note: The format of traj.txt is consistent with that of the replica dataset.

Next, run ENerf-SLAM

python -W ignore run.py configs/Syn/syn1.yaml

Hamlyn Dataset

The data can be found here. Then organize the data as follows.

Datasets/hamlyn
|---results
|      |----depth000000.png
|      |----depth000001.png
|      ...
|      ...
|      |----frame000000.png
|      |----frame000001.png
|      ...
|      ...

Note: There is no groundtruth trajectory for Hamlyn Dataset.

Next, run ENerf-SLAM

python -W ignore run.py configs/Hamlyn/rectified01.yaml

C3VD Dataset

The data can be found here. Then organize the data as follows.

Datasets/desc_t4_a(trans_t2_b)
|---color_undistorted
|      |----0000_color.png
|      |----0001_color.png
|      ...
|      ...
|---depth_undistorted
|      |----0000_depth.tiff
|      |----0001_depth.tiff
|      ...
|      ...
|---pose.txt

Next, run ENerf-SLAM

python -W ignore run.py configs/C3VD/trans2b.yaml

Related Repositories

Our codebase is based on NICE-SLAM and Endo-Depth-and-Motion. We are grateful to the authors for sharing their codebase with the public. Your significant contributions have been instrumental in making our work a reality!

Citation

If you find our code or paper useful, please cite

@article{shan2024enerf,
  title={ENeRF-SLAM: A Dense Endoscopic SLAM With Neural Implicit Representation},
  author={Shan, Jiwei and Li, Yirui and Xie, Ting and Wang, Hesheng},
  journal={IEEE Transactions on Medical Robotics and Bionics},
  year={2024},
  publisher={IEEE}
}