【24h】

Study on Region-Based Forensic Image Retrieval

机译:基于区域的法医图像检索研究

获取原文

摘要

Forensic image database plays an important role in public security area. In order to narrow down the 'semantic gap' between the abundant semantic meanings of the query and the low-level image features, and to improve the precision of the content-based image retrieval (CBIR) for forensic images, this paper proposes a method which combines region semantic template with the ontology describing the relationship between region semantics in an image and the class of the image. In this method a number of objects are selected as representative components of the class. For every object, a set of sample regions are chosen to obtain the semantic template of the object, defined as the average feature of the sample regions. During the retrieval process, images containing same object as that in the query are selected, and then the ontology is used to further region the list to find the desired images. Experimental results prove the proposed method to be effective in forensic image database retrieval.
机译:取证图像数据库在公安领域起着重要作用。为了缩小查询的丰富语义与底层图像特征之间的“语义鸿沟”,并提高取证图像的基于内容的图像检索(CBIR)的准确性,提出了一种方法它结合了区域语义模板和描述图像中区域语义与图像类别之间关系的本体。在这种方法中,选择了许多对象作为类的代表性组件。对于每个对象,选择一组样本区域以获得该对象的语义模板,该模板定义为样本区域的平均特征。在检索过程中,选择包含与查询中相同的对象的图像,然后使用本体对列表进行进一步的区域查找,以找到所需的图像。实验结果证明了该方法在取证图像数据库检索中是有效的。

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