Discriminative archetypal self-taught learning for multispectral landcover classification
Discriminative archetypal self-taught learning for multispectral landcover classification
复制标题
DOI：
10.1109/prrs.2016.7867022
复制
发表时间：
2016-12
期刊：
影响因子：
--
通讯作者：
R. Roscher;Susanne Wenzel;B. Waske
中科院分区：
文献类型：
--
作者：
R. Roscher;Susanne Wenzel;B. Waske
自学学习(STL)已经成为利用未标记数据进行分类的一种很有前途的范例。最常用的自学学习方法是稀疏表示法，其中假设每个样本可以由未标记词典的元素的加权线性组合表示。本文提出了一种用于土地覆盖分类的判别性原型自学习方法，通过选取未标记的判别性原型样本来构建一个功能强大的词典。我们的主要贡献是提出了一种利用可逆跳跃马尔可夫链蒙特卡罗方法来联合确定最佳原型集和构建词典的元素数量的方法。实验使用合成数据、乌克兰一个研究区的多光谱Landsat 7图像和苏黎世基准数据集进行，该数据集包括20幅多光谱Quickbird图像。实验结果表明，该方法能够学习可区分的特征进行分类，与使用原始特征表示的自学习方法相比，以及与随机初始化的原型词典相比，具有更好的分类效果。
Self-taught learning (STL) has become a promising paradigm to exploit unlabeled data for classification. The most commonly used approach to self-taught learning is sparse representation, in which it is assumed that each sample can be represented by a weighted linear combination of elements of a unlabeled dictionary. This paper proposes discriminative archetypal self-taught learning for the application of landcover classification, in which unlabeled discriminative archetypal samples are selected to build a powerful dictionary. Our main contribution is to present an approach which utilizes reversible jump Markov chain Monte Carlo method to jointly determine the best set of archetypes and the number of elements to build the dictionary. Experiments are conducted using synthetic data, a multi-spectral Landsat 7 image of a study area in the Ukraine and the Zurich benchmark data set comprising 20 multispectral Quickbird images. Our results confirm that the proposed approach can learn discriminative features for classification and show better classification results compared to self-taught learning with the original feature representation and compared to randomly initialized archetypal dictionaries.