Localising In Complex Scenes Using Balanced Adversarial Adaptation

TL;DR

This paper introduces a balanced adversarial adaptation method that aligns simulated and real-world representations for indoor localisation, reducing the domain gap without extensive real-world labelling.

Contribution

It proposes a novel symmetrical adversarial approach to improve transferability of localisation representations from simulation to real environments.

Findings

Outperforms fully-supervised methods in real-world indoor localisation

Effectively bridges the domain gap between simulation and real-world data

Maintains geometric invariance while reducing visual discrepancies

Abstract

Domain adaptation and generative modelling have collectively mitigated the expensive nature of data collection and labelling by leveraging the rich abundance of accurate, labelled data in simulation environments. In this work, we study the performance gap that exists between representations optimised for localisation on simulation environments and the application of such representations in a real-world setting. Our method exploits the shared geometric similarities between simulation and real-world environments whilst maintaining invariance towards visual discrepancies. This is achieved by optimising a representation extractor to project both simulated and real representations into a shared representation space. Our method uses a symmetrical adversarial approach which encourages the representation extractor to conceal the domain that features are extracted from and simultaneously preserves robust attributes between source and target domains that are beneficial for localisation. We evaluate our method by adapting representations optimised for indoor Habitat simulated environments (Matterport3D and Replica) to a real-world indoor environment (Active Vision Dataset), showing that it compares favourably against fully-supervised approaches.