文章基本信息

标题：Local Feature Based on Moment Invariants for Blurred Image Matching
本地全文：下载
作者：Qiang Tong ; Terumasa Aoki
期刊名称：International Journal of Electronics and Computer Science Engineering
电子版ISSN：2277-1956
出版年度：2014
卷号：3
期号：4
页码：376-387
出版社：Buldanshahr : IJECSE
摘要：This paper presents a new local feature scheme for image matching between a strongly blurred image and a non-blurred image. In recent years, a lot of local feature schemes have been proposed to improve the image matching performances. However, as far as the authors know, there are no local features which are robust to strong blur. In this paper, blur moment invariants are introduced into a local feature scheme. These blur moment invariants are robust to strong blur when they are used as global features. However, they cannot be used as a local feature. In this paper, we dig into this problem and clarify the reason why they cannot be used as a local feature. After that, we propose a new local feature scheme based on this study. Experimental results show that the proposed scheme is more effective and suitable for blurred image matching than any other existing local feature schemes
关键词：image matching; local feature; moment invariants; de-blurring