Get Your Embedding Space in Order: Domain-Adaptive Regression for Forest Monitoring

TL;DR

This paper introduces a new dataset for cross-domain regression in Earth observation, proposes manifold diffusion as a baseline, and compares various methods under limited data and no prior target domain information.

Contribution

It provides the first dataset for cross-domain regression in remote sensing and evaluates manifold diffusion and other methods for domain adaptation in low-data scenarios.

Findings

Manifold diffusion outperforms other methods in low-data regimes.

Inductive and transductive methods show different advantages in cross-domain regression.

The dataset enables future research in domain-adaptive regression for forest monitoring.

Abstract

Image-level regression is an important task in Earth observation, where visual domain and label shifts are a core challenge hampering generalization. However, cross-domain regression within remote sensing data remains understudied due to the absence of suited datasets. We introduce a new dataset with aerial and satellite imagery in five countries with three forest-related regression tasks. To match real-world applicative interests, we compare methods through a restrictive setup where no prior on the target domain is available during training, and models are adapted with limited information during testing. Building on the assumption that ordered relationships generalize better, we propose manifold diffusion for regression as a strong baseline for transduction in low-data regimes. Our comparison highlights the comparative advantages of inductive and transductive methods in cross-domain regression.