Multi-Task Object Tracking with Feature Selection

Xu CHENG; Nijun LI; Tongchi ZHOU; Zhenyang WU; Lin ZHOU

doi:10.1587/transfun.E98.A.1351

Multi-Task Object Tracking with Feature Selection

Xu CHENG, Nijun LI, Tongchi ZHOU, Zhenyang WU, Lin ZHOU

Full Text Views

0

Share
Cite this

Summary :

In this paper, we propose an efficient tracking method that is formulated as a multi-task reverse sparse representation problem. The proposed method learns the representation of all tasks jointly using a customized APG method within several iterations. In order to reduce the computational complexity, the proposed tracking algorithm starts from a feature selection scheme that chooses suitable number of features from the object and background in the dynamic environment. Based on the selected feature, multiple templates are constructed with a few candidates. The candidate that corresponds to the highest similarity to the object templates is considered as the final tracking result. In addition, we present a template update scheme to capture the appearance changes of the object. At the same time, we keep several earlier templates in the positive template set unchanged to alleviate the drifting problem. Both qualitative and quantitative evaluations demonstrate that the proposed tracking algorithm performs favorably against the state-of-the-art methods.

Publication: IEICE TRANSACTIONS on Fundamentals Vol.E98-A No.6 pp.1351-1354

Publication Date: 2015/06/01

Publicized

Online ISSN: 1745-1337

DOI: 10.1587/transfun.E98.A.1351

Type of Manuscript: LETTER

Category: Image

Authors

Xu CHENG
  Southeast University
Nijun LI
  Southeast University
Tongchi ZHOU
  Southeast University
Zhenyang WU
  Southeast University
Lin ZHOU
  Southeast University

Keyword

visual tracking, sparse representation, multi-task learning, feature selection

Cite this

Copy

Xu CHENG, Nijun LI, Tongchi ZHOU, Zhenyang WU, Lin ZHOU, "Multi-Task Object Tracking with Feature Selection" in IEICE TRANSACTIONS on Fundamentals, vol. E98-A, no. 6, pp. 1351-1354, June 2015, doi: 10.1587/transfun.E98.A.1351.
Abstract: In this paper, we propose an efficient tracking method that is formulated as a multi-task reverse sparse representation problem. The proposed method learns the representation of all tasks jointly using a customized APG method within several iterations. In order to reduce the computational complexity, the proposed tracking algorithm starts from a feature selection scheme that chooses suitable number of features from the object and background in the dynamic environment. Based on the selected feature, multiple templates are constructed with a few candidates. The candidate that corresponds to the highest similarity to the object templates is considered as the final tracking result. In addition, we present a template update scheme to capture the appearance changes of the object. At the same time, we keep several earlier templates in the positive template set unchanged to alleviate the drifting problem. Both qualitative and quantitative evaluations demonstrate that the proposed tracking algorithm performs favorably against the state-of-the-art methods.
URL: https://globals.ieice.org/en_transactions/fundamentals/10.1587/transfun.E98.A.1351/_p

Copy

@ARTICLE{e98-a_6_1351,
author={Xu CHENG, Nijun LI, Tongchi ZHOU, Zhenyang WU, Lin ZHOU, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Multi-Task Object Tracking with Feature Selection},
year={2015},
volume={E98-A},
number={6},
pages={1351-1354},
abstract={In this paper, we propose an efficient tracking method that is formulated as a multi-task reverse sparse representation problem. The proposed method learns the representation of all tasks jointly using a customized APG method within several iterations. In order to reduce the computational complexity, the proposed tracking algorithm starts from a feature selection scheme that chooses suitable number of features from the object and background in the dynamic environment. Based on the selected feature, multiple templates are constructed with a few candidates. The candidate that corresponds to the highest similarity to the object templates is considered as the final tracking result. In addition, we present a template update scheme to capture the appearance changes of the object. At the same time, we keep several earlier templates in the positive template set unchanged to alleviate the drifting problem. Both qualitative and quantitative evaluations demonstrate that the proposed tracking algorithm performs favorably against the state-of-the-art methods.},
keywords={},
doi={10.1587/transfun.E98.A.1351},
ISSN={1745-1337},
month={June},}

Copy

TY - JOUR
TI - Multi-Task Object Tracking with Feature Selection
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1351
EP - 1354
AU - Xu CHENG
AU - Nijun LI
AU - Tongchi ZHOU
AU - Zhenyang WU
AU - Lin ZHOU
PY - 2015
DO - 10.1587/transfun.E98.A.1351
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E98-A
IS - 6
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - June 2015
AB - In this paper, we propose an efficient tracking method that is formulated as a multi-task reverse sparse representation problem. The proposed method learns the representation of all tasks jointly using a customized APG method within several iterations. In order to reduce the computational complexity, the proposed tracking algorithm starts from a feature selection scheme that chooses suitable number of features from the object and background in the dynamic environment. Based on the selected feature, multiple templates are constructed with a few candidates. The candidate that corresponds to the highest similarity to the object templates is considered as the final tracking result. In addition, we present a template update scheme to capture the appearance changes of the object. At the same time, we keep several earlier templates in the positive template set unchanged to alleviate the drifting problem. Both qualitative and quantitative evaluations demonstrate that the proposed tracking algorithm performs favorably against the state-of-the-art methods.
ER -