A Data-Centric Vision Transformer Baseline for SAR Sea Ice Classification

TL;DR

This paper establishes a SAR-only baseline using Vision Transformers for Arctic sea ice classification, demonstrating the effectiveness of focal loss in handling class imbalance and setting a foundation for future multimodal fusion.

Contribution

It introduces a Vision Transformer baseline trained on the AI4Arctic/ASIP dataset with specific techniques, providing a trustworthy benchmark for future multimodal approaches.

Findings

ViT-Large with focal loss achieves 69.6% accuracy on held-out data.

Focal loss improves precision-recall trade-off for rare ice classes.

The baseline offers a cleaner reference for future multimodal fusion research.

Abstract

Accurate and automated sea ice classification is important for climate monitoring and maritime safety in the Arctic. While Synthetic Aperture Radar (SAR) is the operational standard because of its all-weather capability, it remains challenging to distinguish morphologically similar ice classes under severe class imbalance. Rather than claiming a fully validated multimodal system, this paper establishes a trustworthy SAR only baseline that future fusion work can build upon. Using the AI4Arctic/ASIP Sea Ice Dataset (v2), which contains 461 Sentinel-1 scenes matched with expert ice charts, we combine full-resolution Sentinel-1 Extra Wide inputs, leakage-aware stratified patch splitting, SIGRID-3 stage-of-development labels, and training-set normalization to evaluate Vision Transformer baselines. We compare ViT-Base models trained with cross entropy and weighted cross-entropy against a ViT-Large model trained with focal loss. Among the tested configurations, ViT-Large with focal loss achieves 69.6% held-out accuracy, 68.8% weighted F1, and 83.9% precision on the minority Multi-Year Ice class. These results show that focal-loss training offers a more useful precision-recall trade-off than weighted cross-entropy for rare ice classes and establishes a cleaner baseline for future multimodal fusion with optical, thermal, or meteorological data.