Learn Multiple-Kernel SVMs for Domain Adaptation in Hyperspectral Data
Learn Multiple-Kernel SVMs for Domain Adaptation in Hyperspectral Data
复制标题
DOI：
10.1109/lgrs.2012.2236818
复制
发表时间：
2013-09-01
影响因子：
4.8
通讯作者：
Li, Jonathan
中科院分区：
文献类型：
--
作者：
Sun, Zhuo;Wang, Cheng;Li, Jonathan
本文提出了一种新的半监督方法来解决高光谱数据分类中的域自适应问题。为了克服源域和目标域之间分布偏差的影响，我们引入了域转移多核学习来同时最小化支持向量机的最大平均偏差准则和结构风险泛函。然后，将成对二值分类器合并为多类分类器，解决了高光谱数据的分类问题。为了评估该方法对谱分布偏差的稳健性，采用了偏置和非偏置两种采样策略。实际数据集的实验结果表明，该方法在存在跨域分布偏差的情况下仍能达到较高的分类精度，并在不同的标签和未标签数据大小下提供了稳健的解决方案。
This letter presents a novel semisupervised method for addressing a domain adaptation problem in the classification of hyperspectral data. To overcome the influence of distribution bias between the source and target domains, we introduce the domain transfer multiple-kernel learning to simultaneously minimize the maximum mean discrepancy criterion and the structural risk functional of support vector machines. Then, the pairwise binary classifiers are merged as the multiclass classifier for solving the classification problem in hyperspectral data. Both bias and non-bias sampling strategies are introduced to evaluate the robustness of the proposed method against the spectral distribution bias. The results obtained from real data sets show that the proposed method can achieve higher classification accuracy even with cross-domain distribution bias and provide robust solutions with different labeled and unlabeled data sizes.