Finding point correspondence using local similarity and global constraint under insignificant scaling and roll

J. H. Kao*, Jen-Hui Chuang, M. L. Hsieh

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Establishing point correspondences is an important research topic in computer vision. Proposed is an algorithm using local similarity and global constraint to obtain point correspondence. The point correspondences are obtained by comparing their associated colour codes, which are computed by image gradients, and using spatial relationships among neighbouring feature points. The approach assumes insignificant scaling and roll, and is suitable for real-time applications.

Original languageEnglish
Pages (from-to)1340-1342
Number of pages3
JournalElectronics Letters
Volume42
Issue number23
DOIs
StatePublished - 16 Nov 2006

Fingerprint Dive into the research topics of 'Finding point correspondence using local similarity and global constraint under insignificant scaling and roll'. Together they form a unique fingerprint.

Cite this