文章基本信息

标题：新しいフラクタル特徴とロバストな構図情報を用いた画像検索
本地全文：下载
作者：望月貴裕 ; 藤井真人 ; 伊藤崇之等
期刊名称：映像情報メディア学会誌
印刷版ISSN：1342-6907
电子版ISSN：1881-6908
出版年度：2003
卷号：57
期号：6
页码：719-728
DOI：10.3169/itej.57.719
出版社：The Institute of Image Information and Television Engineers
摘要：The rapid increase in the volume of electronically archived images and video materials has given rise to a need for new methods of image retrieval. Therefore, various approaches have been proposed. Most conventional image retrieval methods are based on the content of objects or on features describing only the partial properties of images. But the diverse images we must deal with have induced a necessity to retrieve images more flexibly and effectively, focusing on the multiple properties of images. In this paper, we describe a new method of image retrieval based on two new ideas. One is adopting a fractal sequence designed to include multiple information of texture and to relate one's first impression of an image. The other is structural information of images by allowing rectangular regions, which are set to extract features in multiple points on the image, to be located at slightly varied positions dependent on the image features. Additionally, we show with the results of some experiments that image retrieval that widely perceives the information of images more variously can be achieved because of our applying the new method.
关键词：Active Net;画像検索;テクスチャ特徴;構図情報;フラクタル