LDA单词图像表示的蒙古文古籍图像关键词检索方法
发布时间：2018-10-31 18:56
【摘要】：[目的 ]为了克服传统视觉词袋方法(Bag-of-Visual-Words)中忽略视觉单词间的空间关系和语义信息等问题。[方法 ]本文提出一种与视觉语言模型相结合的基于LDA主题模型,并采用查询似然模型实现检索。[结果 ]实验数据表明,本文所提出的基于LDA的表示方法可以高效、准确地解决蒙古文古籍的关键词检索问题。[结论 ]同时,该方法的性能比Bo VW方法有显著提高。
[Abstract]:[objective] to overcome the problems of ignoring the spatial relationship and semantic information between visual words in traditional visual word bag (Bag-of-Visual-Words). [methods] A subject model based on LDA combined with visual language model is proposed in this paper, and query likelihood model is used to realize retrieval. [results] the experimental data show that the proposed representation method based on LDA can efficiently and accurately solve the keyword retrieval problem of Mongolian ancient books. [conclusion] at the same time, the performance of this method is significantly improved than that of Bo VW method.
【作者单位】： 内蒙古大学图书馆;
【基金】：国家自然科学基金项目“基于领域本体的蒙古文数字资源整合机制研究”(项目编号:71163029)
【分类号】：TP391.3;TP391.41
[Abstract]:[objective] to overcome the problems of ignoring the spatial relationship and semantic information between visual words in traditional visual word bag (Bag-of-Visual-Words). [methods] A subject model based on LDA combined with visual language model is proposed in this paper, and query likelihood model is used to realize retrieval. [results] the experimental data show that the proposed representation method based on LDA can efficiently and accurately solve the keyword retrieval problem of Mongolian ancient books. [conclusion] at the same time, the performance of this method is significantly improved than that of Bo VW method.
【作者单位】： 内蒙古大学图书馆;
【基金】：国家自然科学基金项目“基于领域本体的蒙古文数字资源整合机制研究”(项目编号:71163029)
【分类号】：TP391.3;TP391.41
【相似文献】
相关期刊论文 前10条
1 苏文;王骞;;结合实际案例分析关键词的选取[J];电视技术;2013年S2期
2 黎方正;谢东;;基于完全化语义的关键词检索研究[J];计算机应用研究;2010年10期
3 王，
本文编号：2303225
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2303225.html
