ACGT-Net: Adaptive Cuckoo Refinement-Based Graph Transfer Network for Hyperspectral Image Classification
ACGT-Net: Adaptive Cuckoo Refinement-Based Graph Transfer Network for Hyperspectral Image Classification
复制标题
DOI：
10.1109/tgrs.2023.3307434
复制
发表时间：
2023
影响因子：
8.2
通讯作者：
Yuanchao Su;Jiangyi Chen;Lianru Gao;A. Plaza;Mengying Jiang;Xiang Xu;Xu Sun;Pengfei Li
中科院分区：
文献类型：
--
作者：
Yuanchao Su;Jiangyi Chen;Lianru Gao;A. Plaza;Mengying Jiang;Xiang Xu;Xu Sun;Pengfei Li
深度学习（DL）为高光谱图像分类（HIC）带来了许多新趋势。图神经网络（GNN）是融合DL和结构化数据的模型。虽然基于GNN的方法专注于建模关系，但它们中的大多数都容易受到噪声的影响，不利于从数据中捕获隐藏的相关性。此外，现有的相关方法通常采用固定的图结构，这可能会导致泛化能力差。为了解决上述问题，本文开发了一种自适应布谷鸟细化的图传输网络（ACGT-Net），引入了一种元启发式优化策略来细化图结构。具体来说，我们首先预训练一个图卷积网络（GCN）来学习可转移的权重参数。在无向图中，节点与像素相关联，并且边对应于节点之间的相似性。之后，我们将布谷鸟搜索策略（CSS）集成到训练好的GCN中，以自适应地细化图结构。基于CSS的图结构精化（GSR）算法通过全局优化来关注重要通道，提高GNN的泛化能力。通过对真实的数据集的实验，验证了ACGT-Net与其他SOTA方法相比的有效性和竞争力。
Deep learning (DL) has brought many new trends for hyperspectral image classification (HIC). Graph neural networks (GNNs) are models that fuse DL and structured data. Although GNN-based methods have focused on modeling relations, most of them are susceptible to noise, being adverse to capturing hidden correlations from data. Moreover, the existing related approaches typically adopt changeless graph structures, which might lead to poor generalization. To solve the problems mentioned above, this article develops an adaptive cuckoo refinement-based graph transfer network (ACGT-Net) that introduces a meta-heuristic optimization strategy to refine the graph structure. Specifically, we first pretrain a graph convolutional network (GCN) to learn transferable weight parameters. In the undirected graph, nodes are associated with pixels, and edges correspond to similarities between nodes. Afterward, we integrate a cuckoo search strategy (CSS) into the trained GCN to adaptively refine the graph structure. The graph structure refinement (GSR) with the CSS can pay more attention to significant channels by global optimization to improve the generalization of the GNN. Several experiments with real datasets verify the effectiveness and competitiveness of our ACGT-Net compared with other state-of-the-art (SOTA) methods.