Fine-grained classification aims at distinguishing between items with similar global perception and patterns, but that differ by minute details. Our primary challenges come from both small inter-class variations and large intra-class variations. In this article, we propose to combine several innovations to improve fine-grained classification within the use-case of wildlife, which is of practical interest for experts. We utilize geo-spatiotemporal data to enrich the picture information and further improve the performance. We also investigate state-of-the-art methods for handling the imbalanced data issue.
@article{arxiv.2103.11285,
title = {Geo-Spatiotemporal Features and Shape-Based Prior Knowledge for Fine-grained Imbalanced Data Classification},
author = {Charles A. Kantor and Marta Skreta and Brice Rauby and Léonard Boussioux and Emmanuel Jehanno and Alexandra Luccioni and David Rolnick and Hugues Talbot},
journal= {arXiv preprint arXiv:2103.11285},
year = {2021}
}
Comments
Copyright by the authors. All rights reserved to authors only. Correspondence to: ckantor (at) stanford [dot] edu