Single-View-Place-Recognition

Single-View Place Recognition under Seasonal Changes [arXiv]

Authors: Daniel Olid, Jose M. Facil and Javier Civera

Project Website: [project web]

Partitioned Nordland Dataset: [download]

This work was continued with a Multi-View version of our method, please find our latest work at Condition-Invariant Multi-View Place Recognition

How to use:

We recommend the use of a virtual enviroment for the use of this project. (pew)

$ pew new venvname -p python3 # replace venvname with your prefered name (it also works with python 2.7)

1. Install Caffe

Install Caffe following the instructions.

Please notice if you are using a virtual enviroment to install the python requirements inside the virtual enviroment:

$ pew in venvname
(venvname)$ pip install $req # req variable contains python module to be installed

And add Caffe python module to the PATH in the virtual enviroment:

(venvname)$ pew add caffe/python # replace the path with the path to your caffe repo

2. Install Remaining Dependences

Install remaining dependences:

(venvname)$ for req in $(cat requirements.txt); do pip install $req; done

3. Run Test

To run our demo please run:

(venvname)$ python test_norland.py --help

Note: This version runs with the downsampled version of Partitioned Nordland.

3.1 Download Caffe Models

Fine-tuned version: download

Not Fine-tuned version: download

Contact

You can find my contact information in my Personal Website

License and Citation

This software is under GNU General Public License Version 3 (GPLv3), please see GNU License

For commercial purposes, please contact the authors.

Please cite our paper if it helps your research:

@article{olid2018single,
  title={Single-View Place Recognition under Seasonal Changes},
  author={Olid, Daniel and Facil, Jose M and Civera, Javier},
  journal={arXiv preprint arXiv:1808.06516},
  year={2018}
}

Disclaimer

This site and the code provided here are under active development. Even though we try to only release working high quality code, this version might still contain some issues. Please use it with caution.