Authors: Juan J. Gómez Rodríguez, José M. M. Montiel, Juan D. Tardós.
NR-SLAM is a novel monocular deformable SLAM system founded on the combination of a Dynamic Deformation Graph with a Visco-Elastic deformation model. It is able to reconstruct medical imagery with surfaces with different types of topologies and deformations and can use pinhole and fisheye cameras.
We provide examples to run NR-SLAM in the Hamlyn and in the Endomapper datasets. Videos of some example executions can be found here.
[NR-SLAM] Juan J. Gómez Rodríguez, José M. M. Montiel and Juan D. Tardós, NR-SLAM: Non-Rigid Monocular SLAM, ArXiV xxx.yyy. PDF.
[Deformable tracking] Juan J. Gómez Rodríguez, José M. M. Montiel and Juan D. Tardós, Tracking monocular camera pose and deformation for SLAM inside the human body, IEEE/RSJ International Conference on Intelligent Robots and Systems 2022. PDF.
NR-SLAM is released under AGPL license. For a list of all code/library dependencies (and associated licenses), please see Dependencies.md.
For a closed-source version of NR-SLAM for commercial purposes, please contact the authors: jjgomez (at) unizar (dot) es, josemari (at) unizar (dot) es, tardos (at) unizar (dot) es.
If you use NR-SLAM in an academic work, please cite:
@article{NR-SLAM,
title={{NR-SLAM}: Non-Rigid Monocular {SLAM}},
author={G\´omez, Juan J. AND Montiel,
Jos\'e M. M. AND Tard\'os, Juan D.},
journal={ArXiV xxx.yyy},
year={2023}
}
We have tested the library in Ubuntu 20.04.4 LTS but it should be easy to compile in other platforms. A powerful computer (e.g. i7) will ensure good performance and provide more stable and accurate results.
We use several functionalities of C++17.
We use Pangolin for visualization and user interface. Dowload and install instructions can be found at: https://github.com/stevenlovegrove/Pangolin.
We use OpenCV to manipulate images and features. Dowload and install instructions can be found at: http://opencv.org. Required at leat 3.0. Tested with OpenCV 3.2.0 and 4.4.0.
Required by g2o. Download and install instructions can be found at: http://eigen.tuxfamily.org. Required at least 3.1.0.
We use Boost for directory operations.
We use MLPACK for clustering operations.
Clone the repository:
git clone https://github.com/endomapper/NR-SLAM NR_SLAM
We provide a script build.sh to build the third_party libraries and NR-SLAM. Please make sure you have installed all required dependencies (see section 2). Execute:
cd NR-SLAM
chmod +x build.sh
./build.sh
This will create libNR-SLAM_d at build/lib folder and the executables in build/bin folder.
Endomapper dataset is composed by a set of real and simulated colonoscopies from a monocular endoscope. we provide an example program to launch the sequence for this dataset, both the real and simulated videos.
-
Grab your Endomapper video from https://www.synapse.org/#!Synapse:syn26707219/wiki/615178
-
Execute the following command for a real colonoscopy sequence:
./build/bin/endomapper --dataset_path <video_path>
--settings_path .data/endomapper/settings.yaml
--starting_frame <starting_frame>
--end_frame <last_frame>
- Execute the following command for a simulated colonoscopy sequence:
./build/bin/simulation --dataset_path <dataset_folder>
--settings_path .data/simulation/settings.yaml
--starting_frame <starting_frame>
--end_frame <last_frame>
Hamlyn dataset is a set of endoscopy sequences recorded with a monocular and stereo endoscope.
-
Download the dataset from the webpage. (Disclaimer: unfortunately the web page is often down. Please contact the dataset authors to get the images).
-
Execute the following command:
./build/bin/hamlyn --dataset_path <video_path>
--settings_path .data/hamlyn_<i>/settings.yaml
--starting_frame <starting_frame>
--end_frame <last_frame>