Evaluating Zone-Guided Front Extraction for Glacier Calving-Front Delineation in SAR Imagery

2026-10-02

AI Summary

This study addresses the challenge of accurately extracting glacier calving frontlines in SAR imagery, where boundaries are inherently ambiguous and narrow. Using the CaFFe dataset, we systematically compare direct prediction and region-guided extraction paradigms employing U-Net, DeepLabV3+, and SegFormer-B0 architectures, while evaluating the efficacy of test-time adaptation (TSA). Our results reveal that although region labels encode useful boundary information, frontlines indirectly extracted from model-predicted regions perform significantly worse than those derived from ground-truth regions. Furthermore, TSA yields no substantial improvements. This work exposes the limitations of relying exclusively on region-level evaluation metrics, emphasizing the necessity of independent frontline assessment. We conclude that overcoming current performance bottlenecks requires integrating explicit boundary supervision into future segmentation frameworks.

Abstract

Automatic calving-front delineation from synthetic aperture radar imagery is challenging because the front is a thin and often ambiguous boundary between glacier ice, ocean, and surrounding rock or terrain. The CAlving Fronts and where to Find thEm (CaFFe) dataset provides both binary calving-front masks and broader semantic zone masks, making it possible to study whether zone-level supervision can support front recovery. In this paper, we compare direct front prediction with zone-guided front extraction using U-Net, DeepLabV3+, and SegFormer-B0 under the same bounding-box-cropped CaFFe setting. In the direct setting, models predict the binary calving-front mask. In the zone-guided setting, the model first predicts four semantic zone classes, and the front is then extracted from the predicted glacier-ocean boundary. We evaluate both zone-level and front-level performance, include a ground-truth-zone boundary check, and examine lightweight test-time adaptation on sensor-specific and glacier-specific subsets. The results show that zone labels contain useful front-boundary information: extracting the front from ground-truth zones gives the lowest mean distance error. However, fronts extracted from model-predicted zones remain weak, even when zone segmentation scores are moderate. Test-time adaptation also does not consistently improve zone-guided front recovery. These results indicate that zone segmentation performance should not be treated as a substitute for front-level evaluation and that effective use of zone labels may require boundary-aware training, label fusion, or explicit front supervision.

Problem

• glacier calving-front delineation
• SAR imagery
• zone-guided front extraction
• semantic segmentation
• boundary extraction

Innovation

• Zone-Guided Front Extraction
• Glacier Calving-Front Delineation
• SAR Imagery
• Semantic Segmentation
• Test-Time Adaptation