ConDo: Continual Domain Expansion for Absolute Pose Regression

TL;DR

ConDo introduces a continual learning framework for Absolute Pose Regression that adapts to changing environments by leveraging unlabeled data and knowledge distillation, significantly improving localization accuracy in dynamic scenes.

Contribution

This work presents a novel continual domain expansion method for APR that effectively learns from unlabeled data using knowledge distillation, outperforming standard adaptation techniques.

Findings

Reduces localization error by over 7x in challenging scenes

Outperforms baselines across architectures and scene types

Achieves similar performance to re-training up to 25x faster

Abstract

Visual localization is a fundamental machine learning problem. Absolute Pose Regression (APR) trains a scene-dependent model to efficiently map an input image to the camera pose in a pre-defined scene. However, many applications have continually changing environments, where inference data at novel poses or scene conditions (weather, geometry) appear after deployment. Training APR on a fixed dataset leads to overfitting, making it fail catastrophically on challenging novel data. This work proposes Continual Domain Expansion (ConDo), which continually collects unlabeled inference data to update the deployed APR. Instead of applying standard unsupervised domain adaptation methods which are ineffective for APR, ConDo effectively learns from unlabeled data by distilling knowledge from scene-agnostic localization methods. By sampling data uniformly from historical and newly collected data, ConDo can effectively expand the generalization domain of APR. Large-scale benchmarks with various scene types are constructed to evaluate models under practical (long-term) data changes. ConDo consistently and significantly outperforms baselines across architectures, scene types, and data changes. On challenging scenes (Fig.1), it reduces the localization error by >7x (14.8m vs 1.7m). Analysis shows the robustness of ConDo against compute budgets, replay buffer sizes and teacher prediction noise. Comparing to model re-training, ConDo achieves similar performance up to 25x faster.