Fast Image Matching based on a Multi-scale Description

ASIAIR > College of Computer Science > Department of Computer Science and Information Engineering > Journal Artical > Item 310904400/81350

Please use this identifier to cite or link to this item: http://asiair.asia.edu.tw/ir/handle/310904400/81350

Title:	Fast Image Matching based on a Multi-scale Description
Contributors:	資訊工程學系
Date:	201412
Issue Date:	2014-10-08 06:04:10 (UTC+0)
Abstract:	In order to overcome the Large computation of cross correlation matching, we propose a method of quick cross correlation matching. In traditional cross correlation matching all the pixels take part in computing, so the speed of matching is slow down. In this paper, we use multi-scale edge information which was extracted by improved Laplacian pyramid(ILP) as feather temple; to convert the gray information matching to the feature information matching; take advantage of the feature template and the image to be matched to establish indirect similarity measure to achieve speed-up, anti-geometric distortion. The multi-scale features of the method can meet multiple needs of the matching accuracy and speed. Experiments show, the method improves the speed of cross correlation matching, and has certain robustness.
Relation:	Asia Pacific Signal and Information Processing Association
Appears in Collections:	[Department of Computer Science and Information Engineering] Journal Artical

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	292	View/Open