Unbalanced Optimal Transport Dictionary Learning for Unsupervised Hyperspectral Image Clustering

2026-03-10

AI Summary

This work addresses the challenges of class ambiguity and noise sensitivity in hyperspectral image clustering caused by data imbalance. To this end, we propose a dictionary learning method grounded in unbalanced Wasserstein geometry. By incorporating unbalanced optimal transport and Wasserstein barycenters, our approach circumvents the information loss typically induced by conventional spectral normalization, thereby preserving more faithful spectral structures in low-dimensional representations. The resulting representation is seamlessly integrated with spectral clustering to enable unsupervised segmentation. Experimental results demonstrate that the proposed method significantly enhances clustering sharpness and robustness, effectively maintaining class boundaries and suppressing noise interference without requiring labeled data.

Technology Category

• Machine Learning: Clustering
• Computer Vision: Representation Learning for Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Hyperspectral images capture vast amounts of high-dimensional spectral information about a scene, making labeling an intensive task that is resistant to out-of-the-box statistical methods. Unsupervised learning of clusters allows for automated segmentation of the scene, enabling a more rapid understanding of the image. Partitioning the spectral information contained within the data via dictionary learning in Wasserstein space has proven an effective method for unsupervised clustering. However, this approach requires balancing the spectral profiles of the data, blurring the classes, and sacrificing robustness to outliers and noise. In this paper, we suggest improving this approach by utilizing unbalanced Wasserstein barycenters to learn a lower-dimensional representation of the underlying data. The deployment of spectral clustering on the learned representation results in an effective approach for the unsupervised learning of labels.

Problem

• Unbalanced Optimal Transport
• Dictionary Learning
• Unsupervised Clustering
• Hyperspectral Image
• Wasserstein Barycenters

Innovation

• Unbalanced Optimal Transport
• Dictionary Learning
• Hyperspectral Image Clustering
• Wasserstein Barycenter
• Unsupervised Learning