Nguyen et al. (2025) The Collaborative Border Fuzzy Clustering Approach Based on the Semi-supervised Technique for Land Cover Classification

Identification

• Journal: Lecture notes in networks and systems
• Year: 2025
• Date: 2025-11-12
• Authors: Xuan Phi Nguyen, Dinh Sinh, Nguyễn Long Giang, Trong Hop Dang, Ngoc Cuong Truong, Q.T. Pham
• DOI: 10.1007/978-981-95-1746-6_21

Research Groups

• Graduate University of Sciences and Technology, Vietnam Academy of Science and Technology, Hanoi, Vietnam
• Le Quy Don Technical University, Hanoi, Vietnam
• Hanoi University of Industry, Minh Khai, Bac Tu Liem, Hanoi, Vietnam

Short Summary

This paper introduces a novel semi-supervised border fuzzy clustering approach (B-S²CFC) that combines border fuzzy c-means and semi-supervised collaborative fuzzy clustering for land cover classification from remote sensing imagery, demonstrating superior performance in results and computation time while effectively handling vague cluster boundaries and noise.

Objective

• To develop and evaluate a new fuzzy clustering technique, B-S²CFC, for land cover classification from remote sensing imagery by embedding border and collaborative information into the clustering process to efficiently handle border samples, noise, and outliers.

Study Configuration

• Spatial Scale: Remote sensing imagery, no specific geographic area mentioned.
• Temporal Scale: Not specified, focuses on classification of static imagery.

Methodology and Data

• Models used: Semi-supervised border fuzzy clustering approach based on the collaborative border fuzzy technique (B-S²CFC), which combines Border Fuzzy C-Means (B-FCM) and Semi-supervised Collaborative Fuzzy Clustering (S²CFC).
• Data sources: Remote sensing image data.

Main Results

• The proposed B-S²CFC method significantly outperforms other existing methods in terms of classification results and computational efficiency.
• B-S²CFC effectively addresses the challenges of vague cluster boundaries and noise, which are common issues in remote sensing image classification.
• The approach efficiently solves the problem of border samples within clusters by improving the initial placement of cluster centers based on data distribution, thereby reducing processing time.

Contributions

• Introduction of a novel fuzzy clustering technique (B-S²CFC) that integrates border information and collaborative information into a semi-supervised framework.
• Development of an approach that specifically targets and efficiently handles data samples located at the borders of clusters, which are prone to noise and outliers in remote sensing data.
• Improvement in computational efficiency by optimizing the initial placement of cluster centers.
• Demonstrated effectiveness in mitigating the impact of vague cluster boundaries and noise in land cover classification.

Funding

• Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 105.99-2023.12.

Citation

@article{Nguyen2025Collaborative,
  author = {Nguyen, Xuan Phi and Sinh, Dinh and Giang, Nguyễn Long and Dang, Trong Hop and Truong, Ngoc Cuong and Pham, Q.T.},
  title = {The Collaborative Border Fuzzy Clustering Approach Based on the Semi-supervised Technique for Land Cover Classification},
  journal = {Lecture notes in networks and systems},
  year = {2025},
  doi = {10.1007/978-981-95-1746-6_21},
  url = {https://doi.org/10.1007/978-981-95-1746-6_21}
}