A Novel Semi-Supervised Long-Tailed Learning Framework With Spatial Neighborhood Information for Hyperspectral Image Classification

Feng, Yining; Song, Ruoxi; Ni, Weihan; Zhu, Junheng; Wang, Xianghai

doi:10.1109/lgrs.2023.3241340

Cited by 9 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, several noteworthy works have emerged that focus on utilizing semi-supervised learning techniques for improved hyperspectral image classification and change detection. Within the hyperspectral classification domain, the incorporation of spatial neighborhood information and a semisupervised training framework offers a compelling solution to tackle the issue of imbalanced data distribution [31]. Furthermore, several alternative training strategies, including band selection, active learning and discriminative random field [32]- [34], can be seamlessly integrated into a semisupervised framework to enhance the labeling process and optimize the search for the most relevant spectral bands.…”

Section: A Semi-supervised Learningmentioning

confidence: 99%

Semi-Supervised HyperMatch-Driven Cross Temporal and Spatial Interaction Transformer for Hyperspectral Change Detection

Huang,

Zhang,

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

Hyperspectral images are valuable for precise land cover change detection in a consistent area over time. Nevertheless, supervised methods for hyperspectral change detection face limitations due to insufficient labeled samples. Additionally, current deep learning-based approaches often neglect critical feature interactions across temporal, spatial, and spectral domains. To address these challenges, we introduce a semisupervised model called the HyperMatch-based Cross Temporal and Spatial Interaction Transformer (CTSIT) for hyperspectral change detection. The key contributions of this study are as follows: 1) Introduction of the HyperMatch training schedule, which relies on weak-to-strong consistency, to enhance feature extraction using both labeled and unlabeled data. 2) Contribution of the Cross Temporal Bidirectional Attention (CTBA) module to emphasize bidirectional temporal interactions and resemblances. 3) Introduction of the Spatial and Spectral Attention (SSA) module, which includes the Dense Cross Spatial Attention (DCSA) and Spectral Attention (SA) modules, to capture longrange densely spatial interactions and internal spectral similarities. Extensive comparative experiments conducted on three mainstream hyperspectral change detection datasets confirm the effectiveness and superiority of the proposed HyperMatch-based CTSIT in utilizing both labeled and unlabeled samples. For instance, the training ablation demonstrates that this method significantly outperforms most existing state-of-the-art (SOTA) hyperspectral change detection methods, even with a significantly smaller number of samples.

show abstract

Section: A Semi-supervised Learningmentioning

confidence: 99%

Semi-Supervised HyperMatch-Driven Cross Temporal and Spatial Interaction Transformer for Hyperspectral Change Detection

Huang,

Zhang,

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…At the same time, a multiscale processing mechanism is introduced to obtain different levels of features with graph convolution operations at different scales. Feng et al 33 proposed a semi-supervised long-tail learning framework based on spatial neighborhood information, which can address the HS image classification problem with unbalanced few samples. A new method for inferring labels for unlabeled sample based on spatial neighborhood information improves the accuracy of false labels.…”

Section: Introductionmentioning

confidence: 99%

Multiscale graph convolution residual network for hyperspectral image classification

Li,

Sun,

Feng

et al. 2024

J. Appl. Rem. Sens.

View full text Add to dashboard Cite

In recent years, graph convolutional networks (GCNs) have attracted increased attention in hyperspectral image (HSI) classification through the utilization of data and their connection graph. However, most existing GCN-based methods have two main drawbacks. First, the constructed graph with pixel-level nodes loses many useful spatial information while high computational cost is required due to large graph size. Second, the joint spatial-spectral structure hidden in HSI are not fully explored for better neighbor correlation preservation, which limits the GCN to achieve promising performance on discriminative feature extraction. To address these problems, we propose a multiscale graph convolutional residual network (MSGCRN) for HSI classification. First, to explore the local spatial-spectral structure, superpixel segmentation is performed on the spectral principal component of HSI at different scales. Thus, the obtained multiscale superpixel areas can capture rich spatial texture division. Second, multiple superpixel-level subgraphs are constructed with adaptive weighted node aggregation, which not only effectively reduces the graph size, but also preserves local neighbor correlation in varying subgraph scales. Finally, a graph convolution residual network is designed for multiscale hierarchical features extraction, which are further integrated into the final discriminative features for HSI classification via a diffusion operation. Moreover, a mini-batch branch is adopted to the large-scale superpixel branch of MSGCRN to further reduce computational cost. Extensive experiments on three public HSI datasets demonstrate the advantages of our MSGCRN model compared to several cutting-edge approaches.

show abstract

“…Thus, spatial features are jointly extracted with spectral features, to compensate for information loss and exploit local patterns. The spectral-spatial methods mostly use feature stacking [12], kernel mapping [13][14], filtering [15][16][17], tensor analysis [18][19] and deep learning [20][21][22][23][24], which often come in complicated structures with large computation costs.…”

Section: Introductionmentioning

confidence: 99%

Hybrid Tensor Networks for Fully Supervised and Semisupervised Hyperspectral Image Classification

Xiu,

Ye,

Chen

et al. 2023

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Hyperspectral image (HSI) is rich in spectral information and spatial information to explore the physical and chemical properties of the objects, but it also brings many difficulties to the classification task. The problems of curse of dimensionality and spectral variability in HSIs can affect the efficiency of the classifier and cause the decline of the classification accuracy. Also, the scarcity of manually labeled samples makes it difficult for the fully supervised classifiers to obtain the best results. To this end, this article proposes several novel Hybrid Tensor Networks (HTNs) that represent multiscale spectral-spatial patterns and low-rank features for accurate classification. Moreover, the fully supervised HTN (FHTN) is embedded within a semi-supervised framework with unsupervised modules providing pseudo labels, thus creating semi-supervised HTN (SHTN) to exploit unlabeled data and reduce dependency on manual annotations. With proper post processing techniques, misclassifications are largely reduced and accuracy is further increased. Experimental results show that the proposed HTNs exhibit good generalization and robustness. FHTN and SHTN outperform classic and advanced supervised and semi-supervised models in ground object classification for HSIs.

show abstract

A Novel Semi-Supervised Long-Tailed Learning Framework With Spatial Neighborhood Information for Hyperspectral Image Classification

Cited by 9 publications

References 20 publications

Semi-Supervised HyperMatch-Driven Cross Temporal and Spatial Interaction Transformer for Hyperspectral Change Detection

Semi-Supervised HyperMatch-Driven Cross Temporal and Spatial Interaction Transformer for Hyperspectral Change Detection

Multiscale graph convolution residual network for hyperspectral image classification

Hybrid Tensor Networks for Fully Supervised and Semisupervised Hyperspectral Image Classification

Contact Info

Product

Resources

About