Search papers, labs, and topics across Lattice.
This paper introduces MSCM-net, a novel hyperspectral image classification model that combines multi-scale convolutional networks with Mamba blocks to enhance feature extraction and reduce computational complexity. By integrating a multi-scale feature extraction module and a dual-branch feature aggregation module, the model effectively captures both local and global information, improving classification accuracy. Experimental results on three benchmark datasets demonstrate that MSCM-net outperforms existing methods while maintaining lower computational demands.
MSCM-net achieves superior hyperspectral image classification by leveraging multi-scale convolution and Mamba, striking a balance between accuracy and efficiency.
Hyperspectral imaging is widely used in remote sensing and engineering. Therefore, research on its classification methods is crucial. While CNN and Transformer-based methods have advanced, they still face locality constraints and high computational complexity. To address these issues, we propose an innovative hyperspectral image classification model, MSCM-net. Specifically, first of all, a model architecture combining multi-scale CNN and Mamba is proposed. It consists of a multi-scale feature extraction module (MCSE) and multiple stacked Mamba blocks, which integrates the local feature extraction capability of multi-scale CNN and the long sequence modeling advantage of Mamba. Secondly, the proposed MCSE module consists of multi-scale convolution and SENet. Convolution kernels of different scales extract local information with different receptive fields, enhancing the fusion of spatial and spectral information. Meanwhile, the SENet enables the model to automatically learn the importance of each channel in the multi-scale features. Furthermore, we also propose a dual-branch feature aggregation module, which further effectively extracts and integrates the spectral information contained in the central pixel and the spatial information in the surrounding pixels. Our model has undergone numerous experiments on three widely used benchmark datasets. The experimental results show that MSCM-net can achieve advanced classification performance while reducing computational complexity.
