Compact Multi-level-prior Tensor Representation for Hyperspectral Image Super-resolution

2025-10-07

AI Summary

To address the challenges of jointly modeling multiple priors and high model complexity in hyperspectral image super-resolution, this paper proposes a compact multi-level tensor prior fusion framework. Methodologically: (i) it introduces a block-decomposed higher-order tensor model to decouple spectral low-rankness from spatial structure; (ii) it is the first to jointly model multi-dimensional low-rankness and multi-level spatial total variation within a unified framework, incorporating a non-convex mode-rearranged tensor total variation to enhance higher-order spatial smoothness; and (iii) it employs a linearized alternating direction method of multipliers (ADMM) for optimization, ensuring convergence. Experiments on multiple benchmark datasets demonstrate significant improvements in reconstruction quality, as measured by PSNR and SSIM. The source code is publicly available. The framework achieves a favorable balance between theoretical rigor and practical applicability.

Technology Category

• Machine Learning: Matrix & Tensor Methods
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Fusing a hyperspectral image with a multispectral image acquired over the same scene, extit{i.e.}, hyperspectral image super-resolution, has become a popular computational way to access the latent high-spatial-spectral-resolution image. To date, a variety of fusion methods have been proposed, among which the tensor-based ones have testified that multiple priors, such as multidimensional low-rankness and spatial total variation at multiple levels, effectively drive the fusion process. However, existing tensor-based models can only effectively leverage one or two priors at one or two levels, since simultaneously incorporating multi-level priors inevitably increases model complexity. This introduces challenges in both balancing the weights of different priors and optimizing multi-block structures. Concerning this, we present a novel hyperspectral super-resolution model compactly characterizing these multi-level priors of hyperspectral images within the tensor framework. Firstly, the proposed model decouples the spectral low-rankness and spatial priors by casting the latent high-spatial-spectral-resolution image into spectral subspace and spatial maps via block term decomposition. Secondly, these spatial maps are stacked as the spatial tensor encoding the high-order spatial low-rankness and smoothness priors, which are co-modeled via the proposed non-convex mode-shuffled tensor correlated total variation. Finally, we draw inspiration from the linearized alternating direction method of multipliers to design an efficient algorithm to optimize the resulting model, theoretically proving its Karush-Kuhn-Tucker convergence under mild conditions. Experiments on multiple datasets demonstrate the effectiveness of the proposed algorithm. The code implementation will be available from https://github.com/WongYinJ.

Problem

• Fusing hyperspectral and multispectral images to enhance spatial-spectral resolution
• Overcoming limitations of existing tensor models in handling multi-level priors
• Developing compact tensor representation for hyperspectral image super-resolution

Innovation

• Block term decomposition decouples spectral and spatial priors
• Non-convex mode-shuffled tensor correlated total variation models spatial priors
• Linearized alternating direction method optimizes the model efficiently