An improved algorithm of KPCA-SIFT for image registration

Author

Lu, Xuan-Min ; He, Zhao ; Wang, Jun-Ben

Author_Institution

Sch. of Electron. & Inf., Northwestern Polytech. Univ., Xi´´an, China

fYear

2010

fDate

20-23 June 2010

Firstpage

2493

Lastpage

2496

Abstract

In this paper, an improved SIFT algorithm with Kernel Principal Component Analysis (KPCA-SIFT) is presented for image registration. Gaussian kernel function is applied to the PCA to extract the principal component for reduced-dimension processing of SIFT descriptor to each feature point. The matched keypoints are selected through similar measure, and the Euclidean distance is replaced by linear combination of cityblock and chessboard distances. The experiments show that this algorithm is robust to image changes in scale, noise and rotation with higher matching accuracy.

Keywords

Gaussian processes; image registration; principal component analysis; Euclidean distance; Gaussian kernel function; KPCA-SIFT; Kernel principal component analysis; chessboard distances; cityblock distances; image registration; improved algorithm; Automation; Convolution; Data mining; Euclidean distance; Helium; Image registration; Kernel; Lighting; Noise robustness; Principal component analysis; Image Registration; Kernel PCA; Principal Component Analysis; Scale Invariant Feature transform;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Automation (ICIA), 2010 IEEE International Conference on

Conference_Location

Harbin

Print_ISBN

978-1-4244-5701-4

Type

conf

DOI

10.1109/ICINFA.2010.5512192

Filename

5512192