Contextual HookFormer for Glacier Calving Front Segmentation

Wu F, Gourmelon N, Seehaus T, Zhang J, Braun M, Maier A, Christlein V (2024)

Publication Language: English

Publication Type: Journal article, Original article

Publication year: 2024

Journal

Book Volume: 62

Pages Range: 1-15

Article Number: 5205915

URI: https://ieeexplore.ieee.org/document/10440599

DOI: 10.1109/TGRS.2024.3368215

Abstract

Position changes in glacier calving fronts are important indicators for evaluating the health of ice sheet outlet glaciers and changes in ice dynamics. However, manual delineation of calving fronts in remote sensing imagery is a time-consuming task, resulting in potential large costs. Deep-learning-based methods have made remarkable progress in automatically segmenting and delineating glacier calving fronts from remote sensing imagery. The relatively few remote sensing images and the limited geometric changes for glacier observations both reduce the diversity of the data and exacerbate the difficulty in accurate segmentation. Here we describe a novel automatic method for detecting glacier calving fronts in synthetic aperture radar (SAR) images, termed HookFormer. Our approach processes high-resolution (target) and low-resolution (context) inputs with a unified transformer architecture. The global–local tokens from the context and the target branches are integrated purely by the proposed cross-attention mechanism and cross-interaction module to complement and enhance each other. Moreover, we redesign the HookFormer architecture based on the convolutional neural network (CNN) model AMD-HookNet aiming to improve computational efficiency while achieving significant performance gains with only half of the model parameters/FLOPs. We conduct an in-depth analysis and make extensive comparisons based on the challenging glacier segmentation benchmark dataset CaFFe. As the first pure transformer approach, HookFormer sets a new state of the art with a mean distance error (MDE) of 353m to the ground truth, outperforming the baseline, Swin-Unet, and AMD-HookNet by 53%, 39%, and 19%, respectively.

Authors with CRIS profile

Involved external institutions

University of Chinese Academy of Sciences China (CN)

How to cite

APA:

Wu, F., Gourmelon, N., Seehaus, T., Zhang, J., Braun, M., Maier, A., & Christlein, V. (2024). Contextual HookFormer for Glacier Calving Front Segmentation. IEEE Transactions on Geoscience and Remote Sensing, 62, 1-15. https://doi.org/10.1109/TGRS.2024.3368215

MLA:

Wu, Fei, et al. "Contextual HookFormer for Glacier Calving Front Segmentation." IEEE Transactions on Geoscience and Remote Sensing 62 (2024): 1-15.

BibTeX: Download