Skip to content
Open access

Open-Set Cross-Scene Hyperspectral Image Classification With Collaborative Feature and Classifier Adaptation Networks

2026 · IEEE Transactions on Geoscience and Remote Sensing · Vol 64, pp. 5531814-5531814 · 0 citations · 27 references

TL;DR

A collaborative feature and classifier adaptation network (CF-CAN) that jointly enhances representation robustness and decision reliability under open-set cross-scene conditions and improves both cross-scene generalization and open-set discrimination is proposed.

Abstract

Open-set hyperspectral image classification (HSIC) aims to accurately recognize known categories while rejecting unseen ones. However, existing methods typically assume that training and testing data follow the same distribution. Such an assumption is often violated in cross-scene applications with significant distribution shifts. In this article, we address a more practical and challenging task, termed open-set cross-scene HSIC, which requires simultaneous adaptation to scene distribution discrepancies and reliable detection of unknown categories. To this end, we propose a collaborative feature and classifier adaptation network (CF-CAN) that jointly enhances representation robustness and decision reliability under open-set cross-scene conditions. In particular, an adaptive multiscale feature fusion (AMFF) module is introduced to learn robust and transferable representations, while a class-adaptive dynamic thresholding (CADT) strategy estimates categorywise rejection boundaries for more reliable unknown detection. What’s more, a self-supervised information mining (SSIM) module exploits high-confidence target-domain samples to facilitate domain alignment, improving the separability between known and unknown categories. By collaboratively optimizing feature extraction and classifier adaptation within a unified framework, CF-CAN improves both cross-scene generalization and open-set discrimination. Extensive experiments on three tasks demonstrate that CF-CAN consistently outperforms state-of-the-art HSIC methods.

Read PDF

Similar papers

Open access 2026

Adaptive Boundary Consistency Learning for Open-Set Semi-Supervised Hyperspectral Image Classification

Most existing semi-supervised hyperspectral classification methods are built upon the closed-set assumption, which assumes that unlabeled samples originate exclusively from the known class space. However, this assumption is frequently invalid in real-world remote sensing scenarios. In addition, hyperspectral images (HS...

Hua-Ping Zhou, Shuai Xu, Man Chen et al. · 3 citations
Open access Sep 2026

Frequency-Domain Feature Enhancement Combined with Contrastive Learning for Hyperspectral Image Open-Set Classification

Hyperspectral image classification typically assumes that training and test data belong to the same set of categories. However, in practical applications, unknown categories of ground objects frequently arise during the testing phase, a scenario known as open-set classification. In open-set classification, the high spe...

Cheng-Yang Liu, Hai-Bin Wu, Ai-Li Wang et al. · 0 citations
Open access Oct 2026

Open-Set Hyperspectral Land-Cover Classification Using Second-Order Attention and Gaussian Mixture Clustering

Open-set hyperspectral classification requires accurate recognition of known land-cover classes and detection of samples from unseen classes. We propose GRDOSR, which extends representative-discriminative open-set recognition (RDOSR) by combining second-order attention with Gaussian mixture modeling. Spatial and channe...

Chao-Ya Hu, Ruixue Xv, Mian-Shu Chen · 0 citations
Open access 2026

Adaptive Calibration for Cross-Domain Few-Shot Learning Hyperspectral Image Classification

A dynamic prototype interaction calibration module, which adaptively distinguishes typical samples from hard samples via the prediction confidence of query samples, and a synergistic global domain alignment strategy that jointly aligns macroscopic spatial distributions and spectral baselines while preserving class-disc...

Zhong-Wei Li, Wen-Hui Peng, Qi-Xing Yu et al. · 0 citations
Open access 2026

Hardness-Aware Contrastive Learning With Data Condensation for Open-Set Hyperspectral Image Classification

In recent years, hyperspectral image classification (HSIC) has been widely applied in precision agriculture, environmental monitoring, and remote sensing. However, the frequent emergence of unknown categories, such as novel land-cover types and anomalous regions, poses significant challenges to traditional closed-set c...

Han Wang, Wei Liu, Jing Zhao et al. · 0 citations
Open access 2026

Class-Anchor-Guided Shared Discriminative Learning for Open-Set Hyperspectral-LiDAR Classification

Open-set joint classification of hyperspectral image (HSI) and light detection and ranging (LiDAR) data requires accurate known-class discrimination and reliable rejection of unseen classes. A key challenge is that the fused open-set decision space may be affected by inconsistent class structures learned from modality-...

Ai-Li Wang, Hao-Ran Lv, Hai-Bin Wu et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.