[论文解读] Semi-Supervised Hyperspectral Image Classification with Edge-Aware Superpixel Label Propagation and Adaptive Pseudo-Labeling

Yunfei Qiu, Qiqiong Ma+4 more|arXiv (Cornell University)|Jan 26, 2026

Remote-Sensing Image Classification被引用 0

一句话总结

本论文提出了一种半监督高光谱影像分类框架，将边缘感知超像素标签传播与动态、历史信息驱动的伪标签以及自适应样本分类相结合，以提升伪标签的稳定性与边界处理。

ABSTRACT

Significant progress has been made in semi-supervised hyperspectral image (HSI) classification regarding feature extraction and classification performance. However, due to high annotation costs and limited sample availability, semi-supervised learning still faces challenges such as boundary label diffusion and pseudo-label instability. To address these issues, this paper proposes a novel semi-supervised hyperspectral classification framework integrating spatial prior information with a dynamic learning mechanism. First, we design an Edge-Aware Superpixel Label Propagation (EASLP) module. By integrating edge intensity penalty with neighborhood correction strategy, it mitigates label diffusion from superpixel segmentation while enhancing classification robustness in boundary regions. Second, we introduce a Dynamic History-Fused Prediction (DHP) method. By maintaining historical predictions and dynamically weighting them with current results, DHP smoothens pseudo-label fluctuations and improves temporal consistency and noise resistance. Concurrently, incorporating condifence and consistency measures, the Adaptive Tripartite Sample Categorization (ATSC) strategy implements hierarchical utilization of easy, ambiguous, and hard samples, leading to enhanced pseudo-label quality and learning efficiency. The Dynamic Reliability-Enhanced Pseudo-Label Framework (DREPL), composed of DHP and ATSC, strengthens pseudo-label stability across temporal and sample domains. Through synergizes operation with EASLP, it achieves spatio-temporal consistency optimization. Evaluations on four benchmark datasets demonstrate its capability to maintain superior classification performance.

研究动机与目标

• 解决半监督HSIC中的边界标签扩散和伪标签不稳定性。
• 通过边缘感知超像素传播引入空间先验，提升边界处的标签传播效果。
• 开发伪标签生成与利用的动态学习机制，包括历史融合与三元样本分类。
• 通过统一的DREPL框架将DHP与ATSC结合，提升伪标签的可靠性。
• 在多个基准数据集上评估该方法，以展示鲁棒性与性能提升。

提出的方法

• Edge-Aware Superpixel Label Propagation (EASLP) 将边缘强度惩罚与邻域修正策略相结合，在基于超像素的传播过程中约束标签扩散。
• Dynamic History-Fused Prediction (DHP) 维护每个样本的历史预测，并将历史与当前预测进行融合，历史权重逐步增大。
• Adaptive Tripartite Sample Categorization (ATSC) 使用置信度和 Count-Gap 指标对未标记样本进行易、模糊、难三类分类，并设定自适应阈值。
• Dynamic Reliability-Enhanced Pseudo-Label Framework (DREPL) 将 DHP 与 ATSC 结合，以在时间和样本层面稳定伪标签。
• 使用弱增强与强增强数据扩充，对标记数据使用监督损失，对未标记数据使用自训练损失，伪标签引导。
• 关键方程包括边权相似度 Sim̃ij = Simij / (1 + Ej) 与邻域修正 ˆyj = argmaxc ∑k∈N(j) (1/(Ek+ε)) I[yk=c]。
• research_questions:[

实验结果

研究问题

• RQ1 在HSIs的边缘感知超像素传播中，如何减轻边界的标签扩散？
• RQ2 是否存在基于历史的动态伪标签机制，在训练迭代中稳定伪标签？
• RQ3 未标记样本的自适应分类是否提高伪标签质量与学习效率？
• RQ4 在基准HSIs上，所提出的DREPL框架的总体性能与鲁棒性提升如何？

主要发现

• 所提出的EASLP模块在标签传播中减少边界扩散，同时保持边缘区域的完整性。
• DHP通过利用历史预测并采用动态权重机制，有效平滑伪标签波动。
• ATSC使易、模糊、难样本的选用更具选择性，提升伪标签质量与训练效率。
• DREPL框架将DHP与ATSC结合，提升伪标签在时间与样本层面的可靠性。
• 在四个基准数据集上，与若干全监督、半监督和自监督基线相比，该方法实现了更优的分类性能。