Image Retrieval Method based on Priori Knowledge
首发时间:2012-11-23
Abstract:With the amount of information increasing on the Internet, text search cannot meet the demands of the people, Content-based image retrieval has gradually been familiar to the people. However, its promotion is limited due to the unsatisfactory image search results. Meanwhile, the last retrieval results are affected by the measure and matching methods of local features. In this paper, a measurement method of image features has been presented, we calculate the dimension weight value based on the priori knowledge. Compared with the traditional methods, the experimental results indicate that our method improves the recall and precision ration of image retrieval.
keywords: Content-based image retrieval SURF descriptor Prior knowledge Bag of Words
点击查看论文中文信息
基于先验知识的图像度量方法
摘要:随着互联网上信息量的与日俱增,文本搜索已经不能满足人们的需求,基于内容的图像检索已渐渐为人们所熟悉。但是图像搜索的结果大都不尽如人意,限制了它的推广。其中的局部特征度量方法以及匹配方法对最后检索结果影响较大。本文提出一种基于先验知识的图像特征度量方法,通过该方法同传统的度量方法的检索效果比较,表明该方法在同种条件下能得到更好的检索结果。
关键词: 图像检索 SURF特征 先验知识 Bag of Words
论文图表:
引用
No.****
同行评议
勘误表
基于先验知识的图像度量方法
评论
全部评论0/1000