How Effective is Pre-training of Large Masked Autoencoders for Downstream Earth Observation Tasks?

TL;DR

This paper evaluates the effectiveness of pre-training large Masked Autoencoders for Earth Observation tasks, finding that pre-training benefits are task-dependent and sometimes comparable to training from scratch.

Contribution

It provides a comprehensive analysis of ViT-based MAE pre-training effects on various EO downstream tasks, highlighting when pre-training is advantageous.

Findings

Pre-training benefits are significant for reconstruction tasks.

For segmentation and classification, training from scratch can be equally effective.

Pre-training advantages diminish when downstream tasks differ from pre-training objectives.

Abstract

Self-supervised pre-training has proven highly effective for many computer vision tasks, particularly when labelled data are scarce. In the context of Earth Observation (EO), foundation models and various other Vision Transformer (ViT)-based approaches have been successfully applied for transfer learning to downstream tasks. However, it remains unclear under which conditions pre-trained models offer significant advantages over training from scratch. In this study, we investigate the effectiveness of pre-training ViT-based Masked Autoencoders (MAE) for downstream EO tasks, focusing on reconstruction, segmentation, and classification. We consider two large ViT-based MAE pre-trained models: a foundation model (Prithvi) and SatMAE. We evaluate Prithvi on reconstruction and segmentation-based downstream tasks, and for SatMAE we assess its performance on a classification downstream task. Our findings suggest that pre-training is particularly beneficial when the fine-tuning task closely resembles the pre-training task, e.g. reconstruction. In contrast, for tasks such as segmentation or classification, training from scratch with specific hyperparameter adjustments proved to be equally or more effective.