Superpixel-based hypergraph neural networks with active learning for efficient hyperspectral image classification

Anyong Qin

doi:10.1016/j.neucom.2025.132077

Superpixel-based hypergraph neural networks with active learning for efficient hyperspectral image classification

2025-11-19

0

PRE

AI

DOI：10.1016/j.neucom.2025.132077

Original

Original request for help

Save

Abstract

En 中文

• Novel superpixel-based hypergraph construction strategy enhances HSI spectral-spatial representation by capturing regional spectral-spatial relationships. • Innovative spectral-spatial dual-constraint linking strategy establishes global contextual relationships in hypergraphs. • The SHNN-AL framework introduces superpixel-based hypergraph neural networks with active learning to achieve high classification performance with limited labeled data.

Superpixel-based hypergraph neural networks with active learning for efficient hyperspectral image classification

Abstract

Journal

Organization