M-SIFT: a new descriptor based on Legendre moments and SIFT

Xin Zuo; Xiubin Dai; Limin Luo

doi:10.1117/12.921074

12 January 2012 M-SIFT: a new descriptor based on Legendre moments and SIFT

Xin Zuo, Xiubin Dai, Limin Luo

Proceedings Volume 8350, Fourth International Conference on Machine Vision (ICMV 2011): Computer Vision and Image Analysis; Pattern Recognition and Basic Technologies; 83501B (2012) https://doi.org/10.1117/12.921074
Event: Fourth International Conference on Machine Vision (ICMV 11), 2011, Singapore, Singapore

Abstract

There are many feature descriptors that are insensitive to geometric transformations such as rotation and scale variation. However, most of them cannot effectively deal with blurred image which is a key problem in many real applications. In this paper, we propose a new feature descriptor that combines SIFT descriptor with combined blur, scale and rotation invariant Legendre moment (CBRSL). The proposed method inherits the advantage of SIFT and CBRSL which leads to invariance for scale, rotation and blur degradation simultaneously. We also show how this new descriptor is able to better represent the blur and geometric invariant feature descriptor in image registration. The experimental results validate the effectiveness of our method which is superior to SIFT methods.

Citation Download Citation

Xin Zuo, Xiubin Dai, and Limin Luo "M-SIFT: a new descriptor based on Legendre moments and SIFT", Proc. SPIE 8350, Fourth International Conference on Machine Vision (ICMV 2011): Computer Vision and Image Analysis; Pattern Recognition and Basic Technologies, 83501B (12 January 2012); https://doi.org/10.1117/12.921074

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available