Skip to content

Latest commit

 

History

History
40 lines (29 loc) · 1.64 KB

README.md

File metadata and controls

40 lines (29 loc) · 1.64 KB
Raw

Predicting to Improve: Integrity Measures for Assessing Visual Localization Performance

This repo contains source code for our paper: "Predicting to Improve: Integrity Measures for Assessing Visual Localization Performance", available from the publisher and on QUT ePrints.

Attribution

When using code within this repository, please reference the following paper:

@ARTICLE{9830823,
  author={Carson, Helen and Ford, Jason J. and Milford, Michael},
  journal={IEEE Robotics and Automation Letters}, 
  title={Predicting to Improve: Integrity Measures for Assessing Visual Localization Performance}, 
  year={2022},
  volume={7},
  number={4},
  pages={9627-9634},
  doi={10.1109/LRA.2022.3191205}}

Installation

We recommend using conda or mamba to install all dependencies. Mamba can be installed from mambaforge.

conda create --name vpred_env python=3.9 numpy matplotlib jupyterlab scikit-learn -c conda-forge
conda activate vpred_env

Download the example Nordland feature set using the link here.

Note these features are derived from the partitioned Nordland testset published at http://webdiis.unizar.es/~jmfacil/pr-nordland/#download-dataset by David Olid et al in Single-View Place Recognition under Seasonal Changes In PPNIV Workshop at IROS 2018.

Run the jupyterlab example notebook using:

jupyter lab example.ipynb

Licence

The code is licensed under the MIT License.