WS-Net: Weak-Signal Representation Learning and Gated Abundance Reconstruction for Hyperspectral Unmixing via State-Space and Weak Signal Attention Fusion

2026-03-09

AI Summary

This work addresses the challenge of abundance estimation bias in hyperspectral unmixing, where weak spectral signals are often overwhelmed by dominant endmembers and noise. To mitigate this issue, the authors propose WS-Net, a novel framework that integrates state space models with a weak-signal attention mechanism for the first time. The method employs a multi-resolution wavelet encoder to extract features and combines a Mamba-based state space branch with a dedicated weak-signal attention branch. A learnable gating mechanism adaptively fuses multi-scale information to reconstruct abundance maps, while a KL divergence regularizer explicitly decouples dominant and weak endmembers to alleviate signal collapse. Experiments demonstrate that WS-Net achieves up to 55% lower RMSE and 63% lower SAD on both synthetic and real datasets, maintaining high accuracy and robustness for weak endmembers even under low signal-to-noise ratios.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Intelligent Robots: State Estimation
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Web Mining and Content Analysis: Web data integration and cleaning
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Weak spectral responses in hyperspectral images are often obscured by dominant endmembers and sensor noise, resulting in inaccurate abundance estimation. This paper introduces WS-Net, a deep unmixing framework specifically designed to address weak-signal collapse through state-space modelling and Weak Signal Attention fusion. The network features a multi-resolution wavelet-fused encoder that captures both high-frequency discontinuities and smooth spectral variations with a hybrid backbone that integrates a Mamba state-space branch for efficient long-range dependency modelling. It also incorporates a Weak Signal Attention branch that selectively enhances low-similarity spectral cues. A learnable gating mechanism adaptively fuses both representations, while the decoder leverages KL-divergence-based regularisation to enforce separability between dominant and weak endmembers. Experiments on one simulated and two real datasets (synthetic dataset, Samson, and Apex) demonstrate consistent improvements over six state-of-the-art baselines, achieving up to 55% and 63% reductions in RMSE and SAD, respectively. The framework maintains stable accuracy under low-SNR conditions, particularly for weak endmembers, establishing WS-Net as a robust and computationally efficient benchmark for weak-signal hyperspectral unmixing.

Problem

• weak-signal
• hyperspectral unmixing
• abundance estimation
• spectral response
• endmember

Innovation

• Weak-Signal Attention
• State-Space Modeling
• Hyperspectral Unmixing
• Mamba
• Gated Fusion