Image matching with SIFT descriptor on affine normalized MSERs

Author

Samir, B.V.R. ; Il, Na Sang ; Kalia, Robin

Author_Institution

Contents Protection & Manage. Team, Electron. & Telecommun. Res. Inst. (ETRI), Daejeon, South Korea

fYear

2011

fDate

9-11 Feb. 2011

Firstpage

1

Lastpage

4

Abstract

In this paper, we devise a method for introducing affine invariance into the SIFT algorithm considering the addition of affine invariance property to the efficient SIFT algorithm for image matching. An intermediate step of detecting MSER in Comparison of matching performance was performed with reference to original SIFT algorithm. Other deciding factors such as execution time, number of detected interest points, resolution of the original and target images were also provided for comprehensive comparison. Conclusions are drawn based on the observed results and possible factors are considered for any improvement.

Keywords

content-based retrieval; image matching; SIFT descriptor; affine normalized MSER; content based image retrieval; image matching; Algorithm design and analysis; Computer vision; Detectors; Feature extraction; Image edge detection; Image matching; Pixel; Affine Invariance; Affine Normalization; Image Matching; MSER; SIFT;

fLanguage

English

Publisher

ieee

Conference_Titel

Frontiers of Computer Vision (FCV), 2011 17th Korea-Japan Joint Workshop on

Conference_Location

Ulsan

Print_ISBN

978-1-61284-677-4

Electronic_ISBN

978-1-61284-676-7

Type

conf

DOI

10.1109/FCV.2011.5739761

Filename

5739761