基于多特征融合的图像自动标注精选.pdf

Software Engineering and Applications 软件工程与应用, 2015, 4, 19-25 Published Online April 2015 in Hans. /journal/sea /10.12677/sea.2015.42003 Automatic Image Annotation Based on Multi Feature Fusion Daizhizi Liu, Xunchang Wang, Miao Zang North China University of Technology, Beijing Email: ldzz_ncut@ th th th Received: Apr. 4 , 2015; accepted: Apr. 24 , 2015; published: Apr. 29 , 2015 Copyright © 2015 by authors and Hans Publishers Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). /licenses/by/4.0/ Abstract Automatic image annotation technology has a very important role in terms of retrieving images, and its importance is reflected in the fact that the image can be retrieved in accordance with a certain characteristic image semantic keywords, which brought unprecedented convenience for the retrieval and use of the user. This paper designs a complete set of automatic image annotation technology. Firstly, we extract multiple features from the pictures, and then do the measurements of similarity on the marked images by the method of multi-feature fusions, so that the picture will be labeled automatically. Keywords Multi Feature Fusion, Automatic Annotation, Picture, KNN Algorithm 基于多特征融合的图像自动标注 刘代志子，汪训昌，臧 淼 北方工业大学，北京 Email: ldzz_ncut@ 收稿日期：2015年4月4 日；录用日期：2015年4月24 日；发布日期：2015年4月29 日 摘 要 图像自动标注技术在检索图像方面具有非常重要的作用，其重要性体现在可以根据图像语义关键词来检 19 基于多特征融合的图像自动标注 索具有一定特征的图像，这为用户的检索和使用带来了前所未有的便捷。本文设计了一套完整的图像自 动标注技术，首先对图片的多个特征进行提取，之后通过多特征融合的方法与已标注图像进行相似度量 从而对图片进行自动标注。 关键词 多特征融合，自动标注，图像，KNN算法 1. 引言 随着计算机和现代科技的迅猛发展及广泛应用，数字图像的