Analysis, comparison and improvement of block matching based sub-pixel motion measurement methods for image sequences

Author

Wei, Yangjie ; Dong, Zaili ; Wu, Chengdong

Author_Institution

State Key Lab. of Robot., Grad. Univ. of CAS, Shenyang, China

fYear

2009

fDate

22-24 June 2009

Firstpage

123

Lastpage

127

Abstract

In this paper, the sub-pixel motion measurement for image sequences is researched, and an improved two-dimensional distance-similarity measurement based on block matching algorithm is proposed to decrease the huge computational burden of existing strategies by reducing the searching region. Then, the normal sub-pixel motion estimation based on quadratic curve fitting is analyzed and several other fitting functions are proposed to be used in sub-pixel motion estimation. Finally, the extensively comparison experiments are conducted with respect to the motion measurement for image sequences, and the detailed analysis about influence of the objective functions, the block size, and the fitting function on the estimation precision is given.

Keywords

curve fitting; image sequences; motion estimation; block matching algorithm; fitting function; image sequences; normal subpixel motion estimation; quadratic curve fitting; searching region; subpixel motion measurement methods; two-dimensional distance-similarity measurement; Content addressable storage; Curve fitting; Image analysis; Image motion analysis; Image sequence analysis; Image sequences; Motion estimation; Motion measurement; Pixel; Robotics and automation;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Automation, 2009. ICIA '09. International Conference on

Conference_Location

Zhuhai, Macau

Print_ISBN

978-1-4244-3607-1

Electronic_ISBN

978-1-4244-3608-8

Type

conf

DOI

10.1109/ICINFA.2009.5204905

Filename

5204905