analiza aiqbench
This study evaluates the transferability of computer vision models pretrained on Earth imagery to classify terrain features in Titan's Cassini SAR data, highlighting significant performance variations across different architectures like DINOv2 and DOFA. The findings demonstrate that while pretrained encoders generally outperform classical features, further domain-specific training yields mixed results depending on the model, underscoring the complexities of cross-domain adaptation in remote sensing.
CETUS: How Far Do Representations Trained on Earth Transfer to Cassini SAR of Titan?
arXiv:2610.07576v1 Announce Type: cross Abstract: Cassini synthetic aperture radar (SAR) images reveal the dunes, plains, and lake basins of Titan, providing an instance of representations learned from Earth imagery for planetary terrain classification. Cross-domain Evaluation of Earth-to-Titan Transfer Using SAR (CETUS) compares features from DINOv2, DOFA and CROMA with classical image measurements and features from an untrained vision transformer on the U.S. Geological Survey's Cassini SAR mosaic. The classifiers learn terrain labels from an expert geomorphological map and predict those labels in geographically separate Titan regions. Under logistic regression settings, pretrained encoders achieve higher mean macro F1 than the combined classical features. Encoder rankings change when feature scaling, optimization, and regularization change together. Further training on Titan improves DINOv2 performance, degrades DOFA performance, and leads to mixed results for CROMA under the tested settings. Architectural and input processing differences prevent these comparisons from isolating the effect of pretraining. Classifier fitting and performance on individual terrain classes matter when assessing representation transfer for planetary mapping. Since the map draws partly on the same radar observations, the scores measure agreement with expert interpretation.
Title:CETUS: How Far Do Representations Trained on Earth Transfer to Cassini SAR of Titan?
Abstract:Cassini synthetic aperture radar (SAR) images reveal the dunes, plains, and lake basins of Titan, providing an instance of representations learned from Earth imagery for planetary terrain classification. Cross-domain Evaluation of Earth-to-Titan Transfer Using SAR (CETUS) compares features from DINOv2, DOFA and CROMA with classical image measurements and features from an untrained vision transformer on the U.S. Geological Survey's Cassini SAR mosaic. The classifiers learn terrain labels from an expert geomorphological map and predict those labels in geographically separate Titan regions. Under logistic regression settings, pretrained encoders achieve higher mean macro F1 than the combined classical features. Encoder rankings change when feature scaling, optimization, and regularization change together. Further training on Titan improves DINOv2 performance, degrades DOFA performance, and leads to mixed results for CROMA under the tested settings. Architectural and input processing differences prevent these comparisons from isolating the effect of pretraining. Classifier fitting and performance on individual terrain classes matter when assessing representation transfer for planetary mapping. Since the map draws partly on the same radar observations, the scores measure agreement with expert interpretation.
Research work at NASA Jet Propulsion Laboratory. Available at: this https URL
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
arXiv:2610.07576 [cs.CV]
(or arXiv:2610.07576v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.07576 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 307 cuvinte
Subiecte conexe
Models Can't Remember Their Training. Neither Can You.
LessWrong · 19 iul. 2026
A Minimal Optical-Flow Representation for Vision-Based Tactile Rotation Classification in Robotic Manipulation Across Gravity Domains
arXiv · 9 oct. 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · 13 oct. 2026
BEACON: Behavioral and Semantic Enrichment of AlphaEarth Embeddings through Tri-Modal Contrastive Learning
arXiv · 1 sept. 2026