DCT-Mamba3D: Spectral Decorrelation and Spatial-Spectral Feature Extraction for Hyperspectral Image Classification

2025-02-04

AI Summary

Hyperspectral image classification faces challenges including strong spectral redundancy and complex spatial-spectral dependencies, limiting model robustness and discriminative capability. To address these, we propose an end-to-end framework integrating frequency-domain decorrelation with bidirectional state-space modeling. Specifically, we introduce a novel spectral-spatial decorrelation module based on 3D discrete cosine transform (3D-DCT) to suppress spectral redundancy effectively; design a 3D-Mamba bidirectional state-space model to capture long-range spatial-spectral correlations; and incorporate a global residual enhancement mechanism to improve feature robustness. Extensive experiments on multiple benchmark datasets demonstrate significant improvements over state-of-the-art methods. Notably, our approach achieves substantial accuracy gains in challenging scenarios—such as same-object-different-spectrum and different-object-same-spectrum—validating the effectiveness of synergistic frequency-domain prior guidance and state-space modeling.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Hyperspectral image classification presents challenges due to spectral redundancy and complex spatial-spectral dependencies. This paper proposes a novel framework, DCT-Mamba3D, for hyperspectral image classification. DCT-Mamba3D incorporates: (1) a 3D spectral-spatial decorrelation module that applies 3D discrete cosine transform basis functions to reduce both spectral and spatial redundancy, enhancing feature clarity across dimensions; (2) a 3D-Mamba module that leverages a bidirectional state-space model to capture intricate spatial-spectral dependencies; and (3) a global residual enhancement module that stabilizes feature representation, improving robustness and convergence. Extensive experiments on benchmark datasets show that our DCT-Mamba3D outperforms the state-of-the-art methods in challenging scenarios such as the same object in different spectra and different objects in the same spectra.

Problem

• Reduces spectral redundancy in hyperspectral images
• Captures complex spatial-spectral dependencies
• Enhances feature clarity and robustness

Innovation

• 3D spectral-spatial decorrelation module
• Bidirectional state-space model
• Global residual enhancement module