A hybrid CNN-ViT foundation model trained only on Dutch high-resolution imagery with temporal inputs achieves competitive results on global remote sensing benchmarks despite using fewer parameters and less pretraining data than larger state-of-the-art models.
Rethinking transformers pre-training for multi-spectral satellite imagery
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
Developing a foundation model for high-resolution remote sensing data of the Netherlands
A hybrid CNN-ViT foundation model trained only on Dutch high-resolution imagery with temporal inputs achieves competitive results on global remote sensing benchmarks despite using fewer parameters and less pretraining data than larger state-of-the-art models.