文章基本信息

标题：AdaBoost for Concrete Type of Keywords Annotation
本地全文：下载
作者：Wei-Chao Lin ; Yan-Ze Chen ; Shu-Yuan Chen 等
期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2018
卷号：2233&2234
页码：293-298
出版社：Newswood and International Association of Engineers
摘要：The semantic gap presents an arduous task in semantic-based image retrieval investigations. In this paper, the author proposes the AdaBoost learning algorithm for large vocabulary classification. The main finding of this investigation is that using Gentle AdaBoost for image classification produced excellent results in terms of precision and the F-measure. With 190 concrete keywords categorisation, AdaBoost renders more keywords assignable and allows a significant improvement in all accuracy measures: precision, recall and F-measure. An AdaBoost vs. SVMs comparison showed that AdaBoost was an effective classifier using the one-versus-the-rest mode of operation.
关键词：Image annotation 、 Content;based image retrieval、AdaBoost learning algorithm