Appearance-Invariant 6-DoF Visual Localization using Generative Adversarial Networks

TL;DR

This paper introduces a visual localization method that leverages a CycleGAN-based feature extractor to achieve appearance invariance, enabling accurate 6-DoF localization across different environmental conditions.

Contribution

The paper presents a novel appearance-invariant feature extraction approach using CycleGAN within a localization network, improving long-term outdoor localization under varying weather and seasons.

Findings

Outperforms state-of-the-art methods in challenging environment scenarios

Effective in capturing appearance-invariant features from unpaired samples

Demonstrates robustness across diverse outdoor conditions

Abstract

We propose a novel visual localization network when outside environment has changed such as different illumination, weather and season. The visual localization network is composed of a feature extraction network and pose regression network. The feature extraction network is made up of an encoder network based on the Generative Adversarial Network CycleGAN, which can capture intrinsic appearance-invariant feature maps from unpaired samples of different weathers and seasons. With such an invariant feature, we use a 6-DoF pose regression network to tackle long-term visual localization in the presence of outdoor illumination, weather and season changes. A variety of challenging datasets for place recognition and localization are used to prove our visual localization network, and the results show that our method outperforms state-of-the-art methods in the scenarios with various environment changes.