Investigating Different Geo Priors for Image Classification

TL;DR

This paper evaluates various Spatial Implicit Neural Representations as geographical priors to improve species classification from visual data, analyzing their configurations and handling of unseen species.

Contribution

It introduces a comprehensive evaluation of SINR models as geo priors for species classification, highlighting factors influencing their effectiveness.

Findings

Certain SINR configurations improve classification accuracy.

Handling of unseen species significantly affects model performance.

Factors influencing geo prior effectiveness differ from traditional range maps.

Abstract

Species distribution models encode spatial patterns of species occurrence making them effective priors for vision-based species classification when location information is available. In this study, we evaluate various SINR (Spatial Implicit Neural Representations) models as a geographical prior for visual classification of species from iNaturalist observations. We explore the impact of different model configurations and adjust how we handle predictions for species not included in Geo Prior training. Our analysis reveals factors that contribute to the effectiveness of these models as Geo Priors, factors that may differ from making accurate range maps.