Wu et al. (2026) FCFNet: A Frequency-Domain-Guided Cross-Modal Feature Fusion Network for Cloud Removal

Identification

• Journal: IEEE Transactions on Geoscience and Remote Sensing
• Year: 2026
• Date: 2026-01-01
• Authors: Caifeng Wu, Feng Xu, Xin Li, Fulian Zhao, Zhennan Xu, Xin Lyu
• DOI: 10.1109/tgrs.2026.3662395

Research Groups

[Information not provided in the given text.]

Short Summary

This paper introduces FCFNet, a novel frequency-domain-guided cross-modal feature fusion network specifically designed to address the challenge of cloud removal in imagery.

Objective

• To develop and evaluate FCFNet, a new neural network architecture that leverages frequency-domain guidance and cross-modal feature fusion to effectively remove clouds from images.

Study Configuration

• Spatial Scale: [Specifics not provided, but likely involves image-level or pixel-level processing of remote sensing data.]
• Temporal Scale: [Specifics not provided, but could involve single-image cloud removal or processing of image time series.]

Methodology and Data

• Models used: FCFNet (Frequency-Domain-Guided Cross-Modal Feature Fusion Network).
• Data sources: [Specifics not provided, but typically involves satellite imagery (e.g., optical, synthetic aperture radar (SAR), multispectral) for cloud removal tasks.]

Main Results

[Information not provided in the given text.]

Contributions

• Introduction of FCFNet, a novel network architecture for cloud removal incorporating frequency-domain guidance.
• Development of a cross-modal feature fusion mechanism within FCFNet to enhance cloud removal performance.

Funding

[Information not provided in the given text.]

Citation

@article{Wu2026FCFNet,
  author = {Wu, Caifeng and Xu, Feng and Li, Xin and Zhao, Fulian and Xu, Zhennan and Lyu, Xin},
  title = {FCFNet: A Frequency-Domain-Guided Cross-Modal Feature Fusion Network for Cloud Removal},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  year = {2026},
  doi = {10.1109/tgrs.2026.3662395},
  url = {https://doi.org/10.1109/tgrs.2026.3662395}
}