Class-Aware Consistency Learning for Open-Set Semi-Supervised Hyperspectral Image Classification

doi:10.1109/TGRS.2025.3636993

Class-Aware Consistency Learning for Open-Set Semi-Supervised Hyperspectral Image Classification

2025-01-01

0

PRE

AI

DOI：10.1109/TGRS.2025.3636993

Original

Original request for help

Save

Abstract

En 中文

Semi-supervised hyperspectral image (HSI) classification methods focus on exploring the spectral and spatial information of unlabeled samples. However, existing methods generally follow the closed-set setting, assuming that unlabeled samples do not contain novel classes, which is hard to hold in practical applications. This article aims to study semi-supervised HSI classification in the open-set setting, i.e., unlabeled samples fall into novel classes, and proposes a class-aware consistency learning (CACL) method. First, to explore discriminative spectral–spatial features, a position-aware transformer (PAT) is developed, which effectively models spatial position priors between the center pixel and its neighboring pixels via a symmetric position-aware encoding (PAE). Then, to reduce the interference from novel class samples on the model’s discrimination, a prototype-driven consistency learning (PDCL) is proposed, which accurately selects unlabeled samples belonging to known classes via a known class sampler, and efficiently utilizes their spectral–spatial information by modeling consistent predictions across different views. Finally, to further improve the distinguishability between known classes, a prototype contrastive optimization (PCO) is proposed to decrease the distance between samples from the same class and increase the distances between those from different classes in the feature domain. Furthermore, an adaptive segmentation threshold is designed to accurately predict known classes and reject novel classes. Extensive experiments verify that our CACL outperforms the state-of-the-art methods, achieving the overall accuracy (OA) of 81.65%, 88.61%, and 92.88% on the Indian Pines (IP), Salinas (SA), and Pavia University (PU) datasets, with ten labeled samples in each known class. The code is available at https://github.com/rock-in/CACL-main.

Keywords:

Class prototype

hyperspectral image (HSI) classification

open-set classification

semi-supervised learning (SSL)

Class-Aware Consistency Learning for Open-Set Semi-Supervised Hyperspectral Image Classification

Abstract

Journal

Organization