Hyperspectral image classification, which aims to assign a unique class label to each pixel, has garnered growing attention in recent years. Although many single-source domain adaptation or domain generalization techniques for hyperspectral image classification have been developed, their generalization capability remai...
Pu-Hong Duan, Xu-Dong Zhou, Chuan-Yi Liu et al.· Visual Intelligence· 0 citations
Small-object detection remains highly challenging due to their small size, weak texture, and severe environmental interference, even when both visible and infrared modalities are used, especially in remote sensing images. To address this issue, this article proposes small-object MoE-enhanced spatial prior-guided DiffDe...
Zian Wang, Pu-Hong Duan, Jin Yuan et al.· IEEE Transactions on Geoscie...· 0 citations
A multi-level convolutional collaborative classification network based on the DefMamba fusion mechanism (DMFNet) is proposed in this paper and achieves the highest mean overall accuracy among the compared methods with 10 labeled training samples per class.
Yan Mo, Zi-Yang Liu, Yu-Hao Liu et al.· Remote Sensing· 0 citations
Hyperspectral image (HSI) classification requires effective spectral–spatial representation learning under limited labeled samples and constrained computational resources. Existing Mamba-based methods generally preserve explicit sequence structures through dense or fixed tokens, but this may retain substantial spectral...
Chun-Chao Li, Lu-Lu Shi, Zhuo-Jun Xie et al.· IEEE Journal of Selected Top...· 0 citations
Hyperspectral anomaly detection (HAD) technique, which aims to identify abnormal targets that are different from the surrounding background in both spectral and spatial aspects, plays an important role in diverse remote sensing applications. Recently, a series of deep learning-based methods on the principle of backgrou...
Pu-Hong Duan, Zhi-Xiang Huang, Xu-Dong Kang et al.· IEEE Transactions on Pattern...· 0 citations
A systematic survey and diagnostic evaluation of MLLMs for RSISU and demonstrates the strong transferability of general-purpose CV-MLLMs and shows that current RS-MLLMs do not consistently outperform them across diverse RSISU tasks.