图像检索中相关反馈技术的特性研究  

Characteristic Research of Relevance Feedback Technology in Image Retrieval

作　　者：李云[1] 刘嘉敏[1] 强保华[1] 叶春晓[1] 吴中福[1] 

机构地区：[1]重庆大学计算机学院,重庆400044

出　　处：《计算机工程》2004年第7期128-129,169,共3页Computer Engineering

摘　　要：介绍了基于内容的图像检索技术以及一些图像检索系统存在的局限性，分析了为解决此局限而引入的相关反馈技术的机制，它将人也作为检索系统的一个部分，根据用户的反馈重新查询，从而提高检索精度。 并将相关反馈技术分成3类，研究了采用相关反馈技术的图像检索系统的时序特性。The paper introduces the technique of content-based image retrieval (CBIR) and the limitation of many developed retrieval system, attaches much importance to the mechanism of relevance feedback(RF), which is introduced to break the limitation. The RF looks on the customers as a part of retrieval system and retrieves again according to the feedback information of customers to improve the accuracy of retrieval. And it analyzes three kinds of RF and elaborates on the time sequence characteristic of image retrieval system with RF.

关 键 词：基于内容的图像检索 相关反馈 相似性 时序特性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]