【24h】

FLAGS RETRIEVAL USING COLOR STRINGS

机译:使用颜色字符串检索标志

获取原文
获取原文并翻译 | 示例
           

摘要

The difference between a content-based flag image system and a text-based flag image retrieval system is the capability that the previous one can rank flag images by the degree of similarity between the query flag image and flag images in the database, namely similarity-based retrieval. Conversely, a text-based flag image retrieval system is based on precise match. Our flags retrieval scheme is developed to retrieve flags efficiently. Firstly, we transfer each flag to a color string using only 8 rules automatically. Secondly, we utilize the color strings for comparing the flags, namely color strings comparison (CSC). Our system offers both advantages of the content-based flags retrieval system (similarity-based retrieval) and a text-based flags retrieval one (very rapid and mature). The major differences between our proposed method and existing ones are shown as follows. 1) We consider the spatial information (our color strings comparison is permutation comparison). Most color quantization-based methods are combination comparison. 2) We resized the images to 20*20 pixels. Therefore, our approach can handle diverse sizes, defocus and noise problems. 3) We adopt the concept of the relative relation of RGB. Consequently, our system can deal with dissimilar lighting conditions, partial occlusion, and various color saturation at the same time. We succeed in transferring the flags retrieval problem to the strings comparison. 4) We create a bridge between a content-based retrieval system and a text-based retrieval one. Therefore, the content-based flags retrieval system becomes an analogous text-based retrieval system. Thus, the computational complexity is decreased significantly.
机译:基于内容的标记图像系统和基于文本的标记图像检索系统之间的区别在于,前者可以通过查询标记图像和数据库中标记图像之间的相似度对标记图像进行排名,即基于检索。相反,基于文本的标记图像检索系统基于精确匹配。我们的标志检索方案旨在有效地检索标志。首先,我们仅使用8条规则将每个标志转移到一个颜色字符串。其次,我们利用颜色字符串比较标记,即颜色字符串比较(CSC)。我们的系统既提供了基于内容的标志检索系统(基于相似度的检索)的优点,又提供了基于文本的标志检索系统(非常快速和成熟)的优点。我们提出的方法与现有方法的主要区别如下。 1)我们考虑空间信息(我们的颜色字符串比较是排列比较)。大多数基于颜色量化的方法都是组合比较。 2)我们将图像尺寸调整为20 * 20像素。因此,我们的方法可以处理各种尺寸,散焦和噪声问题。 3)我们采用RGB相对关系的概念。因此,我们的系统可以同时处理不同的照明条件,部分遮挡和各种颜色饱和度。我们成功地将标志检索问题转移到字符串比较中。 4)我们在基于内容的检索系统和基于文本的检索系统之间建立了桥梁。因此,基于内容的标志检索系统变成了类似的基于文本的检索系统。因此,计算复杂度显着降低。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利