An optimized target tracking algorithm of online boosting

Liang Yi Zhu, Rui Zhu, Qing Wang, Dan Lü

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

On the basis of online boosting algorithm, this paper presents an optimized target tracking algorithm. Now, the commonly-used tracking algorithm usually tracks the targets as a whole, thus facing great difficulty in realizing effective tracking while the target is severely shielded. In order to surmount the difficulties resulting from shielding, this paper puts forward block-based target tracking algorithm that splits the target into several parts and then locates the target through summing up the tracking results of every part. To meet the changes of target appearance in tracking process, this paper brings forth block-based target tracking algorithm of online boosting in combination with block-based thinking and online boosting framework. The experimental results illustrate that this algorithm has upgraded tracking robustness and accuracy.

Original languageEnglish
Title of host publication2014 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages450-455
Number of pages6
ISBN (Electronic)9781479952748
DOIs
StatePublished - 15 Dec 2014
Event2014 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2014 - Guilin, Guangxi, China
Duration: 5 Aug 20148 Aug 2014

Publication series

Name2014 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2014

Conference

Conference2014 IEEE International Conference on Signal Processing, Communications and Computing, ICSPCC 2014
Country/TerritoryChina
CityGuilin, Guangxi
Period5/08/148/08/14

Keywords

  • block algorithm
  • online boosting algorithm
  • strong classifier
  • visual tracking

Fingerprint

Dive into the research topics of 'An optimized target tracking algorithm of online boosting'. Together they form a unique fingerprint.

Cite this