Mean shift object tracking with modified LTP

Author

Zhang Shaona ; Hu Dong ; Hu Yanting

Author_Institution

Educ. Minist.´s Key Lab. of Broadband Wireless Commun. & Sensor Network Technol., Nanjing Univ. of Posts & Telecommun., Nanjing, China

Volume

2

fYear

2012

fDate

21-25 Oct. 2012

Firstpage

1139

Lastpage

1143

Abstract

A new scheme of object tracking with mean shift is put forward in this paper. At first, texture feature is fused in the processing by Local Ternary Pattern (LTP). Since LTP is sensitive to local noise, least median of squares (LMedS) algorithm is used to adaptively calculate the noise threshold for accurate estimation of the LTP texture information. Furthermore, target scale and orientation is estimated in case of partial occlusion or rotation, so as to realize robust object tracking. Experimental results show that the proposed algorithm can acquire robust tracking performance under complex background .

Keywords

image texture; least mean squares methods; object tracking; sensor fusion; LMedS; LTP texture information; least median of squares algorithm; local ternary pattern; mean shift object tracking; modified LTP; noise threshold; partial occlusion; robust tracking performance; texture feature; LMedS; LTP; Mean Shift; Object tracking;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing (ICSP), 2012 IEEE 11th International Conference on

Conference_Location

Beijing

ISSN

2164-5221

Print_ISBN

978-1-4673-2196-9

Type

conf

DOI

10.1109/ICoSP.2012.6491779

Filename

6491779