Hyperspectral image (HSI) subspace clustering re- mains a challenging task due to the poor spatial and rich spectral resolutions of HSIs. Most of the existing HSI subspace clustering approaches just extract the spatial and spectral features, ignoring the intrinsic distribution information of data and leading to low accuracy of clustering generally. To solve this problem, this paper presents a diffusion subspace clustering model (DiffSC) which learns distribution information of HSI data simultane- ously through a diffusion module (DM). Specifically, due to the diffusion probabilistic model (DPM) learning raw object data distribution to generate data of the same distribution, which has received wide attention in generation tasks and outperforms other generative models significantly, we attempt to apply the DPM in the field of feature extraction. DiffSC performs distribution information extraction of hyperspectral images by DM and fuses them with spatial-spectral features extracted by deep subspace clustering for training jointly. Experiment outcomes demonstrate that intermediate activation of specific timestep in the inverse diffusion process captures latent distribution information of images effectively and improves the HSI clustering accuracy significantly. Since the DPM is simplified, it can be easily trained from scratch. We evaluate the presented DiffSC on five real HSI datasets, and the experiments indicate that DiffSC can obtain the most advanced clustering outcomes that notably outperform most existing HSI subspace clustering approaches.

The utilization of both optical and synthetic aperture radar (SAR) images for joint semantic segmentation enhances the accuracy of land use classification. Recent advancements in multimodal fusion models, particularly those using self-attention mechanisms and convolutional neural networks (CNNs), have yielded significant results. However, self-attention has quadratic computational complexity, and CNN has insufficient local-global contextual modeling power. Recently, 2-D-selective-scan (SS2D) has emerged as a promising approach. It excels in modeling long-range dependencies while maintaining linear computational complexity. Based on SS2D, we propose a joint semantic segmentation network for optical and SAR images, called MBSSNet. Specifically, we introduce SS2D and design a cross-modal fusion module (CMFM) to fuse multimodal features from dual branches layer by layer, thereby enhancing the consistency of fused feature representations. In addition, during the decoding phase, we integrate contextual information from multiscale fusion features, thereby enhancing the spatial and semantic information of the fused features. Our experimental results show that our method outperforms the state-of-the-art (SOTA), and overall accuracy (OA), mean intersection over union (mIoU), and Kappa outperform other SOTA methods by 1.7%, 3.1%, and 2.2%, respectively.