ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Conference Paper Geometric Method for Detection of Image Quadrilateral
Cited - time in scopus Share share facebook twitter linkedin kakaostory
Authors
Joo-Haeng Lee
Issue Date
2015-06
Citation
International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC) 2015, pp.482-484
Publisher
IEEE
Language
English
Type
Conference Paper
Abstract
A geometric method to detect a quadrilateral area in an image is proposed. The underlying geometric primitive is a line segment represented by two endpoints, which can be detected from an image with an existing detector. The initial set of line segments is filtered with simple heuristics such as weights and length. Given a seed point of a target quadrilateral, a nearest-point measure is applied for further filtering. For the final search set of nearest points, we build Voronoi diagram that manages geometric proximity and accelerates neighborhood search. The result is given as ordered vertices of an image quad in subpixel precision. The proposed method shows an interactive performance for test images.
KSP Keywords
Geometric Primitives, Geometric method, Interactive performance, Line segments, Nearest points, Voronoi Diagram, neighborhood search, seed point