This is the official GitHub page for the paper (Link):
Eric Müller-Budack, Kader Pustu-Iren, Ralph Ewerth: “Geolocation Estimation of Photos using a Hierarchical Model and Scene Classification”. In: European Conference on Computer Vision (ECCV), Munich, Springer, 2018, 575-592.
The source code and trained models previously hosted in this repository have been removed.
This decision was made after careful consideration of the ethical and societal risks associated with image geolocalization technologies, particularly their potential for misuse. In recent years, we observed that our publicly accessible demo https://labs.tib.eu/geoestimation was used in ways that were not aligned with the intended research and educational purposes. As an initial mitigation step, we disabled the analysis of user‑uploaded images. After further reflection, we concluded that withdrawing the implementation entirely is the most responsible course of action.
Our decision is informed by and aligned with concerns raised in recent literature, notably:
Haas, L., Skreta, M., Alberti, S., & Finn, C. (2024).
PIGEON: Predicting Image Geolocations.
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).
https://arxiv.org/abs/2307.05845
As discussed by Haas et al., image geolocalization constitutes a dual‑use technology. While such systems enable beneficial applications—including geographic analysis, education, journalism, environmental monitoring, and autonomous navigation—they also pose privacy, surveillance, and military‑use risks, particularly as predictive accuracy improves. These risks warrant careful scrutiny and proactive mitigation by the research community.
In light of these considerations, we have chosen to:
We encourage researchers and practitioners to continue engaging in responsible AI development, including explicit discussion of downstream impacts and appropriate safeguards for dual‑use technologies.
Status: This repository is maintained for documentation purposes only.
No executable code or trained models are available.
15th December 2025: We decided to delete our models due to ethical concerns that were also pointed out by:
Haas, L., Skreta, M., Alberti, S., & Finn, C. (2024). Pigeon: Predicting image geolocations. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 12893-12902). https://arxiv.org/pdf/2307.05845
12th December 2020 - PyTorch version:
We release a PyTorch implementation and provide weights of a pre-trained base(M, f*) model with underlying ResNet50 architecture.
17th February 2020 - original_tf branch: Since our code is not compatible with TensorFlow 2 and relies on a Caffe model for scene classification, we have added a Dockerfile to simplify the installation. In addition, we have modified the inference script. It is now able to automatically generate the reported results in the paper for the testing datasets.
base(M, f*) model in PyTorch including code for training, validation, testing, and inference.A graphical demonstration where you can compete against the deep learning approach presented in the publication can be found on: https://labs.tib.eu/geoestimation
This demo additionally supports uploading and analyzing your own images. A simplified offline version is located in this repository: https://tibhannover.github.io/GeoEstimation/
We provide the original TensorFlow 1.14 implementation based on a ResNet101 to reproduce the results reported in the paper in the original_tf branch.
We have updated our code in PyTorch including scripts for training, inference and test. We have retrained the base(M, f*) model using a ResNet50 architecture and it achieved comparable results despite its lower complexity.
To use the pre-trained model by default, first download the model checkpoint by running:
mkdir -p models/base_M
wget https://github.com/TIBHannover/GeoEstimation/releases/download/pytorch/epoch.014-val_loss.18.4833.ckpt -O models/base_M/epoch=014-val_loss=18.4833.ckpt
wget https://github.com/TIBHannover/GeoEstimation/releases/download/pytorch/hparams.yaml -O models/base_M/hparams.yaml
Inference with pre-trained model:
python -m classification.inference --image_dir resources/images/im2gps/
Available argparse parameter:
--checkpoint CHECKPOINT
Checkpoint to already trained model (*.ckpt)
--hparams HPARAMS
Path to hparams file (*.yaml) generated during training
--image_dir IMAGE_DIR
Folder containing images. Supported file extensions: (*.jpg, *.jpeg, *.png)
--gpu
Use GPU for inference if CUDA is available, default to true
--batch_size BATCH_SIZE
--num_workers NUM_WORKERS
Number of workers for image loading and pre-processing
Example output that is also stored in a CSV file:
img_id p_key pred_class pred_lat pred_lng
Tokyo_00070_439717934_3d0fd200f1_180_97324495@N00 hierarchy 5367 41.4902 -81.7032
429881745_35a951f032_187_37718182@N00 hierarchy 8009 37.1770 -3.5877
104123223_7410c654ba_19_19355699@N00 hierarchy 7284 32.7337 -117.1520
The (list of) image files for testing can be found on the following links:
Download and extract the two testsets (Im2GPS, Im2GPS3k) in resources/images/<dataset_name> and run the evaluation script with the provided meta data, i.e., the ground-truth coordinate for each image.
When using the default paramters, make sure that the pre-trained model is available.
# download im2gps testset
mkdir resources/images/im2gps
wget http://graphics.cs.cmu.edu/projects/im2gps/gps_query_imgs.zip -O resources/images/im2gps.zip
unzip resources/images/im2gps.zip -d resources/images/im2gps/
wget https://raw.githubusercontent.com/TIBHannover/GeoEstimation/original_tf/meta/im2gps_places365.csv -O resources/images/im2gps_places365.csv
wget https://raw.githubusercontent.com/TIBHannover/GeoEstimation/original_tf/meta/im2gps3k_places365.csv -O resources/images/im2gps3k_places365.csv
python -m classification.test
Available argparse paramters:
--checkpoint CHECKPOINT
Checkpoint to already trained model (*.ckpt)
--hparams HPARAMS
Path to hparams file (*.yaml) generated during training
--image_dirs IMAGE_DIRS [IMAGE_DIRS ...]
Whitespace separated list of image folders to evaluate
--meta_files META_FILES [META_FILES ...]
Whitespace separated list of respective meta data (ground-truth GPS positions). Required columns: IMG_ID,LAT,LON
--gpu
Use GPU for inference if CUDA is available, default to True
--precision PRECISION
Full precision (32), half precision (16)
--batch_size BATCH_SIZE
--num_workers NUM_WORKERS
Number of workers for image loading and pre-processing
Results on the Im2GPS and Im2GPS3k test sets: The reported accuracies (in percentage) is the fraction of images localized within the given radius (in km) using the GCD distance. Note, that we used the full MP-16 training dataset and all 25600 images for validation, thus the results will differ when not all images are available.
Im2GPS: Model | 1 | 25 | 200 | 750 | 2500 | |:—————|—–:|—–:|——:|——:|——-: | base(M, c) | 9.3 | 31.6 | 49.8 | 67.1 | 78.9 | | base(M, m) | 13.9 | 34.6 | 48.1 | 68.4 | 79.3 | | base(M, f) | 15.6 | 39.2 | 48.9 | 65.8 | 78.5 | | base(M, f)__ | 14.8 | 37.6 | 48.9 | 68.4 | 78.9 | | __base(M, f) (original) | 15.2 | 40.9 | 51.5 | 65.4 | 78.5 |
Im2GPS3k:
| Model | 1 | 25 | 200 | 750 | 2500 |
|---|---|---|---|---|---|
| base(M, c) | 6.2 | 24.3 | 36.3 | 51.7 | 67.0 |
| base(M, m) | 8.3 | 26.2 | 35.7 | 51.4 | 66.5 |
| base(M, f) | 9.9 | 27.3 | 36.2 | 51.2 | 66.4 |
| base(M, f*) | 10.1 | 28.0 | 36.9 | 51.1 | 67.0 |
| base(M, f*) (original) | 9.7 | 27.0 | 35.6 | 49.2 | 66.0 |
| ISN(M, f*, S3) (original) | 10.5 | 28.0 | 36.6 | 49.7 | 66.0 |
We provide a complete training script which is written in PyTorch Lightning and report all hyper-paramters used for the provided model. Furthermore, a script is given to download and pre-process the images that are used for training and validiation.
1) Download training and validation images
- We provide a script to download the images given a list of URLs
- Due to no longer publicly available images, the size of the dataset might be smaller than the original.
- We also store the images in chunks using MessagePack to speed-up the training process (similar to multiple TFRecord files)
2) Given multiple s2 partitionings (e.g. coarse, middle, fine from the paper), the respective classes are assigned to each image on both datasets.
3) Training and hyper-paramters: All hyper-paramters can be configured in configs/baseM.yml as well as paramters from PyTorch Lightning Trainer class.
Necessary steps:
# download and preprocess images
wget https://github.com/TIBHannover/GeoEstimation/releases/download/v1.0/mp16_urls.csv -O resources/mp16_urls.csv
wget https://github.com/TIBHannover/GeoEstimation/releases/download/pytorch/yfcc25600_urls.csv -O resources/yfcc25600_urls.csv
python download_images.py --output resources/images/mp16 --url_csv resources/mp16_urls.csv --shuffle
python download_images.py --output resources/images/yfcc25600 --url_csv resources/yfcc25600_urls.csv --shuffle --size_suffix ""
# assign cell(s) for each image using the original meta information
wget https://github.com/TIBHannover/GeoEstimation/releases/download/v1.0/mp16_places365.csv -O resources/mp16_places365.csv
wget https://github.com/TIBHannover/GeoEstimation/releases/download/pytorch/yfcc25600_places365.csv -O resources/yfcc25600_places365.csv
python partitioning/assign_classes.py
# remove images that were not downloaded
python filter_by_downloaded_images.py
# train geo model from scratch
python -m classification.train_base --config config/baseM.yml
The geographical cell labels are extracted using the S2 geometry library and can be visualized on http://s2.sidewalklabs.com/regioncoverer/. Create a partitioning using the following command for a given dataset (as CSV file) which contains an image id, latitude and longitude. We provide the partitionings that are used in the paper below. ```shell script python partitioning/create_cells.py [-h] [-v] –dataset DATASET –output OUTPUT –img_min IMG_MIN –img_max IMG_MAX [–lvl_min LVL_MIN] [–lvl_max LVL_MAX]
### Requirements
All requirements are listed in the `environment.yml`. We recomment to use [*conda*](https://docs.conda.io/en/latest/) to install all required packages in an individual environment.
```bash
# clone this repo
git clone https://github.com/TIBHannover/GeoEstimation.git && cd GeoEstimation
# install dependencies
conda env create -f environment.yml
conda activate geoestimation-github-pytorch
# download pre-calculated parititonings
mkdir -p resources/s2_cells
wget https://raw.githubusercontent.com/TIBHannover/GeoEstimation/original_tf/geo-cells/cells_50_5000.csv -O resources/s2_cells/cells_50_5000.csv
wget https://raw.githubusercontent.com/TIBHannover/GeoEstimation/original_tf/geo-cells/cells_50_2000.csv -O resources/s2_cells/cells_50_2000.csv
wget https://raw.githubusercontent.com/TIBHannover/GeoEstimation/original_tf/geo-cells/cells_50_1000.csv -O resources/s2_cells/cells_50_1000.csv
@inproceedings{muller2018geolocation,
author = {Müller-Budack, Eric and Pustu-Iren, Kader and Ewerth, Ralph},
title = {Geolocation Estimation of Photos Using a Hierarchical Model and Scene
Classification},
booktitle = {Computer Vision - {ECCV} 2018 - 15th European Conference, Munich,
Germany, September 8-14, 2018, Proceedings, Part {XII}},
series = {Lecture Notes in Computer Science},
volume = {11216},
pages = {575--592},
publisher = {Springer},
year = {2018},
url = {https://doi.org/10.1007/978-3-030-01258-8\_35},
doi = {10.1007/978-3-030-01258-8\_35},
}
This work is published under the GNU GENERAL PUBLIC LICENSE Version 3, 29 June 2007. For details please check the LICENSE file in the repository.