• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Computer Applications in Industry and Engineering >Text Location in Complex Image using Ouster-Based Templates
【24h】

Text Location in Complex Image using Ouster-Based Templates

机译：使用基于Ouster的模板在复杂图像中的文本位置

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The focus of this paper is to find the text region in images containing the multi-segment character such as Korean. Korean has the complex structure and composition rules which makes it difficult to distinguish character-components from background or noise, using only color and size information. For this, we propose to use of shape information as well as color and size information. To represent the shape information, cluster-based templates are obtained from target characters using K-means clustering algorithm. The proposed method consists of three components: image segmentation, noise filtering, and text region detection. After segmenting an image into connected components with uniform color and size, the non-character components are removed by some heuristics. Then, to accurately discriminate character-components from noise, the segmented results are filtered by matching with cluster-based templates, as well as color and size information. And then text region is located from the profile analysis of the character-components. To assess the validity of the proposed method, the experiments are performed with a variety of images such as videos, web images, scanned images.
机译：本文的重点是在包含多段字符（例如韩语）的图像中找到文本区域。韩语具有复杂的结构和构图规则，因此仅使用颜色和大小信息就很难区分字符成分与背景或噪点。为此，我们建议使用形状信息以及颜色和尺寸信息。为了表示形状信息，使用K均值聚类算法从目标字符中获取基于聚类的模板。所提出的方法包括三个部分：图像分割，噪声过滤和文本区域检测。在将图像分割为颜色和大小均一的连接组件后，通过一些启发式方法将非字符组件删除。然后，为了准确地将字符成分与噪声区分开，通过与基于聚类的模板以及颜色和大小信息进行匹配，对分割结果进行了过滤。然后从字符组件的轮廓分析中找到文本区域。为了评估所提出方法的有效性，对各种图像（例如视频，Web图像，扫描图像）进行了实验。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号