Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 4, 2026Open Access

Superpixel-Tokenized and Frequency-Modulated Hybrid CNN–Transformer for Remote Sensing Semantic Segmentation

View Full Paper
Ask AI
Bookmark
Share

Authors

XXXinlin XieCentral South UniversityCCChenhao ChangTaiyuan University of Science and TechnologyYYYunyun YangTaiyuan University of Technology

Discussion

Loading...

Member takes

Implication

Presents a novel method for enhancing semantic segmentation in remote sensing, suggesting improved structural integrity and boundary precision.

Key Points

  • The research aims to improve remote sensing semantic segmentation by integrating superpixel tokens and frequency modulation to enhance boundary accuracy and structural integrity.
  • Developed SFCT-Net integrating superpixel tokens and frequency constraints
  • Implemented a Superpixel-Tokenized Linear Position Attention module for preserving object integrity
  • Constructed a Frequency-Modulated Deformable Edge Refinement module for boundary recovery
  • Created a Spatial–Semantic Feature Coupling module for correcting spatial drift and aligning features
  • SFCT-Net outperformed existing hybrid segmentation architectures in urban scene understanding
  • Achieved robust structural recovery and boundary precision in various datasets
  • Demonstrated significant improvements in performance metrics on Taiyuan Satellite Remote Sensing Dataset and benchmark datasets

Cite This Study

Xie et al. (2026) studied this question.

synapsesocial.com/papers/69a7cd3dd48f933b5eed96b2https://doi.org/10.3390/rs18050754
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Channel reconstruction and dual attention dynamic fusion for remote sensing image semantic segmentation2026 · 1 citations
  2. 2SFA-Net: Semantic Feature Adjustment Network for Remote Sensing Image Segmentation2024 · 39 citations
  3. 3A Transformer-based multi-modal fusion network for semantic segmentation of high-resolution remote sensing imagery2024 · 40 citations
  4. 4CMTFNet: CNN and Multiscale Transformer Fusion Network for Remote-Sensing Image Semantic Segmentation2023 · 360 citations
  5. 5SF3Net: Frequency-Domain Enhanced Segmentation Network for High-Resolution Remote Sensing Imagery2025