Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2504.06962 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915261330751488 |
|---|---|
| author | Kerdreux, Thomas Tuel, Alexandre Febvre, Quentin Mouche, Alexis Chapron, Bertrand |
| author_facet | Kerdreux, Thomas Tuel, Alexandre Febvre, Quentin Mouche, Alexis Chapron, Bertrand |
| contents | Self-supervised learning (SSL) has enabled the development of vision foundation models for Earth Observation (EO), demonstrating strong transferability across diverse remote sensing tasks. While prior work has focused on network architectures and training strategies, the role of dataset curation, especially in balancing and diversifying pre-training datasets, remains underexplored. In EO, this challenge is amplified by the redundancy and heavy-tailed distributions common in satellite imagery, which can lead to biased representations and inefficient training.
In this work, we propose a dynamic dataset pruning strategy designed to improve SSL pre-training by maximizing dataset diversity and balance. Our method iteratively refines the training set without requiring a pre-existing feature extractor, making it well-suited for domains where curated datasets are limited or unavailable. We demonstrate our approach on the Sentinel-1 Wave Mode (WV) Synthetic Aperture Radar (SAR) archive, a challenging dataset dominated by ocean observations. We train models from scratch on the entire Sentinel-1 WV archive spanning 10 years. Across three downstream tasks, our results show that dynamic pruning improves both computational efficiency and representation quality, leading to stronger transferability.
We also release the weights of OceanSAR-1, the first model in the OceanSAR family, a series of foundation models for ocean observation and analysis using SAR imagery, at github.com/galeio-research/OceanSAR-models/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_06962 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation Kerdreux, Thomas Tuel, Alexandre Febvre, Quentin Mouche, Alexis Chapron, Bertrand Computer Vision and Pattern Recognition Artificial Intelligence Self-supervised learning (SSL) has enabled the development of vision foundation models for Earth Observation (EO), demonstrating strong transferability across diverse remote sensing tasks. While prior work has focused on network architectures and training strategies, the role of dataset curation, especially in balancing and diversifying pre-training datasets, remains underexplored. In EO, this challenge is amplified by the redundancy and heavy-tailed distributions common in satellite imagery, which can lead to biased representations and inefficient training. In this work, we propose a dynamic dataset pruning strategy designed to improve SSL pre-training by maximizing dataset diversity and balance. Our method iteratively refines the training set without requiring a pre-existing feature extractor, making it well-suited for domains where curated datasets are limited or unavailable. We demonstrate our approach on the Sentinel-1 Wave Mode (WV) Synthetic Aperture Radar (SAR) archive, a challenging dataset dominated by ocean observations. We train models from scratch on the entire Sentinel-1 WV archive spanning 10 years. Across three downstream tasks, our results show that dynamic pruning improves both computational efficiency and representation quality, leading to stronger transferability. We also release the weights of OceanSAR-1, the first model in the OceanSAR family, a series of foundation models for ocean observation and analysis using SAR imagery, at github.com/galeio-research/OceanSAR-models/. |
| title | Efficient Self-Supervised Learning for Earth Observation via Dynamic Dataset Curation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2504.06962 |