Skip to content
/ cadloc Public

Benchmark for visual localization on imperfect 3D mesh models from the Internet

License

Notifications You must be signed in to change notification settings

v-pnk/cadloc

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Benchmark for visual localization on 3D mesh models

Teaser image Repository for the localization benchmark at v-pnk.github.io/cadloc. To upload the results to the benchmark, run the evaluation script on your pose estimates and create a pull request to add your results into the tables.

Citation

If you use the benchmark or the code from this repository, please cite the following paper:

@article{Panek2023cadloc,
title={{Visual Localization using Imperfect 3D Models from the Internet}},
author={Vojtech Panek and Zuzana Kukelova and Torsten Sattler},
booktitle={Proceedings of the IEEE/CVF conference on computer vision and pattern recognition},
year={2023}}

Installation

The evaluation tool and rendering script needs few Python packages:

conda create --name cadloc python=3.9
conda activate cadloc
pip3 install open3d==0.17 pycolmap==0.3 numpy argparse tqdm Pillow

Preparation of data for the benchmark

The evaluation script expects the estimated poses to be stored in the .txt format used by visuallocalization.net, i.e., each line contains an image name and estimated pose described by a quaternion q and translation vector t:

<name> <q[W]> <q[X]> <q[Y]> <q[Z]> <t[X]> <t[Y]> <t[Z]>

Uploading new benchmark results

Edit the table corresponding to your results and create a pull request.

License

This repository is licensed under the 3-Clause BSD License. See the LICENSE file for full text.

About

Benchmark for visual localization on imperfect 3D mesh models from the Internet

Topics

Resources

License

Stars

Watchers

Forks

Languages