CFormer: Cross-Frequency Transformer for Remote Sensing Semantic Segmentation
IEEE Transactions on Geoscience and Remote Sensing · 2026 · 64 · 1-13
DOI: 10.1109/tgrs.2026.3673568 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Geoscience and Remote Sensing, 2026. Verfasst von Yong Wang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 51 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Research Start-Up Fund of Aerospace Information Research Institute, Henan Academy of Sciences, Scientific Research Foundation of the Henan Academy of Sciences.
Abstract (Originalsprache)
Remote sensing semantic segmentation aims at fine-grained pixel-level understanding of high-resolution imagery, but remains challenging due to large-scale variations, complex textures, and uneven spatial distributions. Existing convolutional neural network (CNN) and transformer-based methods struggle to jointly model global context and local details. To address this issue, we propose a cross-frequency transformer for remote sensing semantic segmentation. The framework introduces a multiscale transformer with frequency-aware channel attention, which explicitly models low-frequency structural information and high-frequency details via discrete cosine transform (DCT), and adaptively fuses frequency-enhanced features to improve boundary preservation and cross-scale discrimination. Furthermore, a Fourier relation transformer is designed in the decoding stage to compute attention in the frequency domain, effectively strengthening long-range dependencies while suppressing high-frequency noise. Extensive experiments on six benchmark datasets demonstrate that the proposed method consistently outperforms existing approaches in overall accuracy (OA), boundary quality, and cross-scale object recognition.
Förderung
- National Natural Science Foundation of China
- Research Start-Up Fund of Aerospace Information Research Institute, Henan Academy of Sciences
- Scientific Research Foundation of the Henan Academy of Sciences
Zitieren
APA
Wang, Y., Li, Y., Wu, R., Zhang, W., An, L., & Geng, G. (2026). CFormer: Cross-Frequency Transformer for Remote Sensing Semantic Segmentation. IEEE Transactions on Geoscience and Remote Sensing, 64, 1-13. https://doi.org/10.1109/tgrs.2026.3673568
Vancouver
Wang Y, Li Y, Wu R, Zhang W, An L, Geng G. CFormer: Cross-Frequency Transformer for Remote Sensing Semantic Segmentation. IEEE Transactions on Geoscience and Remote Sensing. 2026;64:1-13. doi:10.1109/tgrs.2026.3673568
BibTeX
@article{Wang2026,
author = {Wang, Yong and Li, Yuanyuan and Wu, Renzhe and Zhang, Wensong and An, Li and Geng, Guohua},
title = {CFormer: Cross-Frequency Transformer for Remote Sensing Semantic Segmentation},
journal = {IEEE Transactions on Geoscience and Remote Sensing},
year = {2026},
volume = {64},
pages = {1--13},
doi = {10.1109/tgrs.2026.3673568}
}
EUVIMED