Hierarchical Attention Fusion for Geo-Localization

TL;DR

This paper introduces a hierarchical attention fusion network that leverages multi-scale features from CNNs to improve robustness in geo-localization tasks, especially under drastic scale variations.

Contribution

It proposes a novel hierarchical attention fusion approach with self-supervised training for enhanced multi-scale feature integration in geo-localization.

Findings

Outperforms existing state-of-the-art methods on large-scale benchmarks.

Effective handling of drastic scale variations in scene localization.

Self-supervised adaptive weighting improves feature emphasis.

Abstract

Geo-localization is a critical task in computer vision. In this work, we cast the geo-localization as a 2D image retrieval task. Current state-of-the-art methods for 2D geo-localization are not robust to locate a scene with drastic scale variations because they only exploit features from one semantic level for image representations. To address this limitation, we introduce a hierarchical attention fusion network using multi-scale features for geo-localization. We extract the hierarchical feature maps from a convolutional neural network (CNN) and organically fuse the extracted features for image representations. Our training is self-supervised using adaptive weights to control the attention of feature emphasis from each hierarchical level. Evaluation results on the image retrieval and the large-scale geo-localization benchmarks indicate that our method outperforms the existing state-of-the-art methods. Code is available here: \url{https://github.com/YanLiqi/HAF}.