EarthShift is a new benchmark using paired datasets to measure robustness of geospatial foundation models to realistic distribution shifts, finding consistent 15-20% performance drops out-of-distribution across 8 models and 11 tasks.
Geocrossbench: Cross-band generalization for remote sensing.arXiv preprint arXiv:2511.02831, 2025
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CV 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
DeluluNet enables continued prediction under modality substitution, addition, or subsets by training a multi-modal model from a unimodal teacher and unlabeled multimodal data via modality hallucination.
A geometry-faithful RPC-consistent protocol for satellite multi-view features shows semantic similarity does not guarantee geometric matchability and that 2D backbones remain competitive with 3D-aware models under proper constraints.
citing papers explorer
-
EarthShift: a benchmark for measuring robustness to real-world distribution shifts in Earth observation
EarthShift is a new benchmark using paired datasets to measure robustness of geospatial foundation models to realistic distribution shifts, finding consistent 15-20% performance drops out-of-distribution across 8 models and 11 tasks.
-
Changing Modalities: Adapting Remote Sensing Models to New Satellites and Sensors
DeluluNet enables continued prediction under modality substitution, addition, or subsets by training a multi-modal model from a unimodal teacher and unlabeled multimodal data via modality hallucination.
-
Geometric Consistency Protocol for Foundation Model Features in Multi-View Satellite Imagery
A geometry-faithful RPC-consistent protocol for satellite multi-view features shows semantic similarity does not guarantee geometric matchability and that 2D backbones remain competitive with 3D-aware models under proper constraints.