A Bayesian hierarchical model for learning natural scene categories
A Bayesian hierarchical model for learning natural scene categories
复制标题
DOI：
10.1109/cvpr.2005.16
复制
发表时间：
2005-06
期刊：
影响因子：
--
通讯作者：
Li Fei-Fei-Li-Fei-Fei-48004138;P. Perona
中科院分区：
文献类型：
--
作者：
Li Fei-Fei-Li-Fei-Fei-48004138;P. Perona
我们提出了一种新的方法来学习和识别自然场景类别。与以前的工作不同，它不需要专家注释训练集。我们表示一个场景的图像的局部区域的集合，表示为通过无监督学习获得的码字。每个区域都是“主题”的一部分。在以前的工作中，这些主题是从专家的手工注释中学习的，而我们的方法学习主题分布以及主题上的码字分布而无需监督。我们报告令人满意的分类性能上的一个大的13类复杂的场景。
We propose a novel approach to learn and recognize natural scene categories. Unlike previous work, it does not require experts to annotate the training set. We represent the image of a scene by a collection of local regions, denoted as codewords obtained by unsupervised learning. Each region is represented as part of a "theme". In previous work, such themes were learnt from hand-annotations of experts, while our method learns the theme distributions as well as the codewords distribution over the themes without supervision. We report satisfactory categorization performances on a large set of 13 categories of complex scenes.