Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Regional biases in image geolocation estimation: a case study with the SenseCity Africa dataset

Ximena Salgado Uribe, Martí Bosch, Jérôme Chenal

TL;DR

This work applies a state-of-the-art image geolocation estimation model (ISNs) to a crowd-sourced dataset of geolocated images from the African continent (SCA100), and explores the regional and socioeconomic biases underlying the model's predictions.

Abstract

Advances in Artificial Intelligence are challenged by the biases rooted in the datasets used to train the models. In image geolocation estimation, models are mostly trained using data from specific geographic regions, notably the Western world, and as a result, they may struggle to comprehend the complexities of underrepresented regions. To assess this issue, we apply a state-of-the-art image geolocation estimation model (ISNs) to a crowd-sourced dataset of geolocated images from the African continent (SCA100), and then explore the regional and socioeconomic biases underlying the model's predictions. Our findings show that the ISNs model tends to over-predict image locations in high-income countries of the Western world, which is consistent with the geographic distribution of its training data, i.e., the IM2GPS3k dataset. Accordingly, when compared to the IM2GPS3k benchmark, the accuracy of the ISNs model notably decreases at all scales. Additionally, we cluster images of the SCA100 dataset based on how accurately they are predicted by the ISNs model and show the model's difficulties in correctly predicting the locations of images in low income regions, especially in Sub-Saharan Africa. Therefore, our results suggest that using IM2GPS3k as a training set and benchmark for image geolocation estimation and other computer vision models overlooks its potential application in the African context.

Regional biases in image geolocation estimation: a case study with the SenseCity Africa dataset

TL;DR

This work applies a state-of-the-art image geolocation estimation model (ISNs) to a crowd-sourced dataset of geolocated images from the African continent (SCA100), and explores the regional and socioeconomic biases underlying the model's predictions.

Abstract

Advances in Artificial Intelligence are challenged by the biases rooted in the datasets used to train the models. In image geolocation estimation, models are mostly trained using data from specific geographic regions, notably the Western world, and as a result, they may struggle to comprehend the complexities of underrepresented regions. To assess this issue, we apply a state-of-the-art image geolocation estimation model (ISNs) to a crowd-sourced dataset of geolocated images from the African continent (SCA100), and then explore the regional and socioeconomic biases underlying the model's predictions. Our findings show that the ISNs model tends to over-predict image locations in high-income countries of the Western world, which is consistent with the geographic distribution of its training data, i.e., the IM2GPS3k dataset. Accordingly, when compared to the IM2GPS3k benchmark, the accuracy of the ISNs model notably decreases at all scales. Additionally, we cluster images of the SCA100 dataset based on how accurately they are predicted by the ISNs model and show the model's difficulties in correctly predicting the locations of images in low income regions, especially in Sub-Saharan Africa. Therefore, our results suggest that using IM2GPS3k as a training set and benchmark for image geolocation estimation and other computer vision models overlooks its potential application in the African context.
Paper Structure (12 sections, 5 figures, 1 table)

This paper contains 12 sections, 5 figures, 1 table.

Table of Contents

  1. Abstract
  2. Introduction
  3. Background
  4. Materials and Methods
  5. Experiments and results
  6. Discussion and conclusion
  7. Acknowledgments

Figures (5)

  • Figure 1: Example images from the IM2GPS \ref{['fig:im2gps']} and SCA100 \ref{['fig:sca100']} datasets.
  • Figure 2: Contextual geographic divisions of the world. The map on the left (\ref{['fig:income-groups']}) illustrates the delineation of income groups according to the World Bank Group country classifications. The one on the right (\ref{['fig:world-regions']}) shows the division of the world into nine distinct regions based on Jones' classification system.
  • Figure 3: Confusion matrices by world regions (\ref{['fig:confusion-region']} and income groups (\ref{['fig:confusion-income']}) for the IM2GPS3k (left) and SCA100 (right) datasets. The bars over each axis show the row (predicted) and column (observed) totals.
  • Figure 4: Clustegram diagram (\ref{['fig:clustegram']}) based on the distance $d$ between the predicted and the ground truth location in the SCA100 dataset. Diagram obtained using the Python library clustergram fleischmann2023clustergram. Violin plot (\ref{['fig:distance-cluster-violin']}) of the obtained clusters.
  • Figure 5: Confusion matrices for the SCA100 clusters C$_i$ to C$_{iv}$ (top to bottom), for income groups (\ref{['fig:cluster-confusion-income']}) and world regions (\ref{['fig:cluster-confusion-region']}).