Combi-CAM: A Novel Grad-CAM Approach for Geolocalization Explainability

Official repository for Combi-CAM: A Novel Grad-CAM Approach for Geolocalization Explainability paper.

Inference with explainability methods:

!python inference.py --image_url "..." --use_cpu --layercam --gradcam --gradcamplusplus --scorecam --combicam

Recommended to generate figures one by one to avoid RAM overload.

It is also possible to use --image_path instead of --image_url.

Copyright

This repository uses the geolocation model developed by the CERTH team. All rights and credits for the geolocation model belong to CERTH.

Model Weights

The model weights are proprietary to CERTH and will not be released. For access to the weights, please reach out to the authors of [1].

References

[1] Giorgos Kordopatis-Zilos, Panagiotis Galopoulos, Symeon Papadopoulos, and Ioannis Kompatsiaris. 2021. Leveraging EfficientNet and Contrastive Learning for Accurate Global-scale Location Estimation. In Proceedings of the 2021 International Conference on Multimedia Retrieval (ICMR '21). Association for Computing Machinery, New York, NY, USA, 155–163. https://doi.org/10.1145/3460426.3463644