Combi-CAM: A Novel Multi-Layer Approach for Explainable Image Geolocalization
Authors: David Faget, José Luis Lisani, Miguel Colom
Organizations: CB · Centre Borelli, ENS Paris-Saclay, Université de Paris, CNRS, INSERM, SSA, France · CB, CMLA · Universitat de les Illes Balears, IAC3, Spain
Planet-scale photo geolocalization involves the intricate task of estimating the geographic location depicted in an image purely based on its visual features. While deep learning models, particularly convolutional neural networks (CNNs), have significantly advanced this field, understanding the reasoning behind their predictions remains challenging. In this paper, we present Combi-CAM, a novel method that enhances the explainability of CNN-based geolocalization models by combining gradient-weighted class activation maps obtained from several layers of the network architecture, rather than using only information from the deepest layer as is typically done. This approach provides a more detailed understanding of how different image features contribute to the model's decisions, offering deeper insights than the traditional approaches.
Figures & tables
Figure 1: The Eiffel Tower and its replicas: Original in Paris (left), and replicas in Las Vegas (center) and Madrid (right).
Figure 2: Aerial image of Paris (France) showing a wide outlook of the city. It includes characteristic objects such as the Eiffel Tower, the Seine river, and buildings with particular architectural elements.
Figure 3: Results obtained by applying Grad-CAM on the last layer of selected blocks (from #0 to #31) of the EfficientNet-B4 architecture show that characteristic elements, such as the Eiffel Tower, become more prominent in the middle blocks of the analysis compared to the final blocks.
Figure 4: The maximum activation magnitude per block indicates that the highest activations occur in blocks #22 to #29, where the pixels of characteristic elements activate the most. More specifically, the most significant block is the 22 which prominently highlights the most recognizable element of Paris: the Eiffel Tower.
Figure 5: View of Sydney (Australia).
Figure 6: Activation maps obtained with different methods for the images in Figures 2 and 5 . Left column, Paris’ image. Right column, Sydney’s image.
Figure 7: Activation maps obtained with different methods for the images in Figure 1 . Left column, Eiffel tower in Paris; center, replica in Las Vegas; right, replica in Madrid.
When humans play geolocation games such as GeoGuessr, they rely on concrete visual cues, such as road markings, vegetation, or architectural details, to infer where an image was captured. Whether image geolocation models rely on similar object-level evidence remains difficult to determine, as attribution methods like Grad-CAM typically highlight diffuse regions rather than coherent visual entities, making it difficult to link model predictions to specific objects or perceptible patterns. In this work, we propose an object-centric analysis pipeline to investigate the visual evidence used by geolocation models. Starting from attribution maps, we extract salient regions and segment them into object-like elements. We evaluate their predictive relevance through deletion and insertion tests, comparing attributionguided crops to randomly selected regions with similar coverage. Experiments on a three-country benchmark show that attribution-guided crops consistently retain more information for the model's prediction than random crops. These results suggest that attribution maps can be decomposed into interpretable, perceptible elements, providing a step toward object-level analysis of geolocation models.
Emilie Durrieu, Christophe Hurter, Philippe Muller +1
ENAC, University of Toulouse · IRIT, University of Toulouse, ANITI · CNRS
Worldwide image geo-localization aims to determine where on Earth a single image was captured. However, visually similar scenes may lie thousands of kilometers apart, so methods that localize primarily by appearance often mistake a distant look-alike for the true location. We attribute this failure to a structural cause: in existing methods, GPS coordinates serve only as training supervision, and the distance relationships among locations never enter the learned representation. To address this, we propose GeoMetric, a retrieval-based framework that encodes GPS coordinates relationally rather than in isolation, injecting the distance structure among locations into both representation learning and inference. GeoMetric comprises three components: (1) a Transformer-based GPS encoder with distance-aware location attention that modulates inter-sample aggregation by great-circle proximity; (2) a trimodal contrastive objective that aligns images, geo-textual descriptions, and GPS embeddings in a unified space; and (3) a retrieval-augmented inference stage that supplies large multimodal models (LMMs) with contrastive candidate context for grounded coordinate reasoning. Extensive experiments on IM2GPS, IM2GPS3k, YFCC4k, and YFCC26k demonstrate that GeoMetric consistently outperforms state-of-the-art methods, improving street-level accuracy (within 1 km) by 1.5%, 0.9%, 6.9%, and 2.5%, respectively. Controlled ablations confirm that the gains originate from the proposed geographic encoding rather than any specific LMM.
Junchao Cui, Xuanzi Ma, Wenqi Shi +3
Information Engineering University Zhengzhou, China
Image geolocation aims to infer the geographic origin of an image from visual content alone. However, this task remains challenging in regions where countries share similar urban, roadside, architectural, and environmental characteristics. Many existing geolocation models focus on coordinate level prediction or classification performance while providing limited insight into how visual evidence contributes to location predictions. This study presents an explainable country level image geolocation pipeline for 11 ASEAN countries. First, we collected 4,850 images from GeoGuessr style sources, Google Images, and additional street level imagery. We then evaluated three approaches on this dataset: CLIP zero shot classification, a LightGBM classifier, and an MLP classifier. The MLP achieved the best test performance, attaining an accuracy and F1 score of 85.91%. For explainability, predictions generated by the MLP classifier were analyzed post hoc using CLIP attention rollout, YOLO26 object detection on the original images, and Energy Based Pointing Game (EBPG) overlap metrics. Object level analysis indicates that frequently detected objects are not necessarily associated with the highest attention density, suggesting that object frequency and attention based visual evidence capture different aspects of a scene. These results demonstrate that the proposed model can support accurate regional image geolocation while enabling object level inspection of the visual cues underlying its predictions.