Robust Superpixel Tracking with Weighted Multiple-Instance Learning

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

doi:10.1587/transinf.2014EDL8176

Robust Superpixel Tracking with Weighted Multiple-Instance Learning

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

Full Text Views

0

Share
Cite this

Summary :

This paper proposes a robust superpixel-based tracker via multiple-instance learning, which exploits the importance of instances and mid-level features captured by superpixels for object tracking. We first present a superpixels-based appearance model, which is able to compute the confidences of the object and background. Most importantly, we introduce the sample importance into multiple-instance learning (MIL) procedure to improve the performance of tracking. The importance for each instance in the positive bag is defined by accumulating the confidence of all the pixels within the corresponding instance. Furthermore, our tracker can help recover the object from the drifting scene using the appearance model based on superpixels when the drift occurs. We retain the first (k-1) frames' information during the updating process to alleviate drift to some extent. To evaluate the effectiveness of the proposed tracker, six video sequences of different challenging situations are tested. The comparison results demonstrate that the proposed tracker has more robust and accurate performance than six ones representing the state-of-the-art.

Publication: IEICE TRANSACTIONS on Information Vol.E98-D No.4 pp.980-984

Publication Date: 2015/04/01

Publicized: 2015/01/15

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2014EDL8176

Type of Manuscript: LETTER

Category: Image Recognition, Computer Vision

Authors

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

Keyword

visual tracking, multiple instance learning, appearance model, superpixel

Cite this

Copy

Xu CHENG, Nijun LI, Tongchi ZHOU, Lin ZHOU, Zhenyang WU, "Robust Superpixel Tracking with Weighted Multiple-Instance Learning" in IEICE TRANSACTIONS on Information, vol. E98-D, no. 4, pp. 980-984, April 2015, doi: 10.1587/transinf.2014EDL8176.
Abstract: This paper proposes a robust superpixel-based tracker via multiple-instance learning, which exploits the importance of instances and mid-level features captured by superpixels for object tracking. We first present a superpixels-based appearance model, which is able to compute the confidences of the object and background. Most importantly, we introduce the sample importance into multiple-instance learning (MIL) procedure to improve the performance of tracking. The importance for each instance in the positive bag is defined by accumulating the confidence of all the pixels within the corresponding instance. Furthermore, our tracker can help recover the object from the drifting scene using the appearance model based on superpixels when the drift occurs. We retain the first (k-1) frames' information during the updating process to alleviate drift to some extent. To evaluate the effectiveness of the proposed tracker, six video sequences of different challenging situations are tested. The comparison results demonstrate that the proposed tracker has more robust and accurate performance than six ones representing the state-of-the-art.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2014EDL8176/_p

Copy

@ARTICLE{e98-d_4_980,
author={Xu CHENG, Nijun LI, Tongchi ZHOU, Lin ZHOU, Zhenyang WU, },
journal={IEICE TRANSACTIONS on Information},
title={Robust Superpixel Tracking with Weighted Multiple-Instance Learning},
year={2015},
volume={E98-D},
number={4},
pages={980-984},
abstract={This paper proposes a robust superpixel-based tracker via multiple-instance learning, which exploits the importance of instances and mid-level features captured by superpixels for object tracking. We first present a superpixels-based appearance model, which is able to compute the confidences of the object and background. Most importantly, we introduce the sample importance into multiple-instance learning (MIL) procedure to improve the performance of tracking. The importance for each instance in the positive bag is defined by accumulating the confidence of all the pixels within the corresponding instance. Furthermore, our tracker can help recover the object from the drifting scene using the appearance model based on superpixels when the drift occurs. We retain the first (k-1) frames' information during the updating process to alleviate drift to some extent. To evaluate the effectiveness of the proposed tracker, six video sequences of different challenging situations are tested. The comparison results demonstrate that the proposed tracker has more robust and accurate performance than six ones representing the state-of-the-art.},
keywords={},
doi={10.1587/transinf.2014EDL8176},
ISSN={1745-1361},
month={April},}

Copy

TY - JOUR
TI - Robust Superpixel Tracking with Weighted Multiple-Instance Learning
T2 - IEICE TRANSACTIONS on Information
SP - 980
EP - 984
AU - Xu CHENG
AU - Nijun LI
AU - Tongchi ZHOU
AU - Lin ZHOU
AU - Zhenyang WU
PY - 2015
DO - 10.1587/transinf.2014EDL8176
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E98-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2015
AB - This paper proposes a robust superpixel-based tracker via multiple-instance learning, which exploits the importance of instances and mid-level features captured by superpixels for object tracking. We first present a superpixels-based appearance model, which is able to compute the confidences of the object and background. Most importantly, we introduce the sample importance into multiple-instance learning (MIL) procedure to improve the performance of tracking. The importance for each instance in the positive bag is defined by accumulating the confidence of all the pixels within the corresponding instance. Furthermore, our tracker can help recover the object from the drifting scene using the appearance model based on superpixels when the drift occurs. We retain the first (k-1) frames' information during the updating process to alleviate drift to some extent. To evaluate the effectiveness of the proposed tracker, six video sequences of different challenging situations are tested. The comparison results demonstrate that the proposed tracker has more robust and accurate performance than six ones representing the state-of-the-art.
ER -