Zone labels offer clues but do not replace direct glacier front detection
Evaluating Zone-Guided Front Extraction for Glacier Calving-Front Delineation in SAR Imagery
Computer Vision and Pattern RecognitionMachine Learning
Summary
Finding the exact edge where glaciers meet the ocean in radar images is hard because the boundary is thin and unclear. The authors studied two ways to find this edge: one predicts the edge directly, and the other predicts surrounding zones before extracting the edge. They found that knowing the zones helps if the zones are perfectly labeled, but when the zones are predicted by a computer, the edge detection is still poor. This means just good zone predictions aren't enough, and special techniques are needed to improve edge detection.
What this means in practice
- •For environmental monitoring teams: Use semantic zone labels to improve glacier calving-front boundary identification for better ice-ocean interaction tracking in SAR images.
- •For satellite imagery analysts: Incorporate zone-level annotations to enhance glacier front delineation workflows despite the need for further refinement of boundary detection methods.
Authors
Chhaya Kulkarni, Emam Hossain
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.