Feature Extraction Using Weighted Training Samples
Feature Extraction Using Weighted Training Samples
复制标题
DOI：
10.1109/lgrs.2015.2402167
复制
发表时间：
2015-07-01
影响因子：
4.8
通讯作者：
Ghassemian, Hassan
中科院分区：
文献类型：
--
作者：
Imani, Maryam;Ghassemian, Hassan
提出了一种基于加权训练样本的特征提取方法。不同的光谱波段(特征)在土地覆盖分类识别中起着不同的作用。在FEWT中，获得训练样本的每个特征在预测该样本的类别标签中的相对重要性，并将其作为该特征的权重。然后，加权后的训练样本可以用于每种任意的特征提取方法。在本文中，我们将加权训练样本用于有监督的局部保持投影。在三幅高光谱图像上的实验结果表明，FEWT在训练样本数量有限的情况下，比现有的一些有监督的特征提取方法具有更好的性能和更快的速度。
Feature extraction using weighted training (FEWT) samples is proposed in this letter. Different spectral bands (features) play different roles in identification of land-cover classes. In the FEWT, the relative importance of each feature of a training sample in predicting the class label of that sample is obtained and considered as a weight for that feature. Then, the weighted training samples can be used in each arbitrary feature extraction method. In this letter, we use the weighted training samples in supervised locality preserving projection. The experimental results on three popular hyperspectral images show that FEWT has better performance and more speed than some state-of-the-art supervised feature extraction methods using limited number of available training samples.