An Improved Methodology for Image Feature Matching

This paper presents a novel methodology to perform matching between image points described by their respective features. Traditionally, such correspondences are determined by computing the similarity between descriptor vectors associated to each point which are obtained by invariant descriptors. Our...

Full description

Saved in:
Bibliographic Details
Published in2009 XXII Brazilian Symposium on Computer Graphics and Image Processing pp. 307 - 314
Main Authors da Camara Neto, V.F., Campos, M.F.M.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2009
Subjects
Computational efficiency
Computational geometry
Computer graphics
Computer vision
Euclidean distance
Feature Correspondence
Image matching
Image processing
Layout
Lighting
Photometry
SIFT
SURF
Online AccessGet full text
ISBN1424449782
9781424449781
ISSN1530-1834
DOI10.1109/SIBGRAPI.2009.24

Cover

More Information
Summary:This paper presents a novel methodology to perform matching between image points described by their respective features. Traditionally, such correspondences are determined by computing the similarity between descriptor vectors associated to each point which are obtained by invariant descriptors. Our methodology first obtains a coarse global registration among images, which constrains the correspondence space. Then, it analyzes the similarity among descriptors, thus reducing both the number and the severity of mismatches. The approach is sufficiently generic to be used with many feature descriptor methods. We present several experimental results that show significant increase in accuracy, number of successful matches, and execution time.
ISBN:1424449782
9781424449781
ISSN:1530-1834
DOI:10.1109/SIBGRAPI.2009.24