IEICE TRANSACTIONS on Information
Siamese Visual Tracking with Dual-Pipeline Correlated Fusion Network
Summary :
Siamese visual tracking, viewed as a problem of max-similarity matching to the target template, has absorbed increasing attention in computer vision. However, it is a challenge for current Siamese trackers that the demands of balance between accuracy in real-time tracking and robustness in long-time tracking are hard to meet. This work proposes a new Siamese based tracker with a dual-pipeline correlated fusion network (named as ADF-SiamRPN), which consists of one initial template for robust correlation, and the other transient template with the ability of adaptive feature optimal selection for accurate correlation. By the promotion from the learnable correlation-response fusion network afterwards, we are in pursuit of the synthetical improvement of tracking performance. To compare the performance of ADF-SiamRPN with state-of-the-art trackers, we conduct lots of experiments on benchmarks like OTB100, UAV123, VOT2016, VOT2018, GOT-10k, LaSOT and TrackingNet. The experimental results of tracking demonstrate that ADF-SiamRPN outperforms all the compared trackers and achieves the best balance between accuracy and robustness.
- Publication
- IEICE TRANSACTIONS on Information Vol.E104-D No.10 pp.1702-1711
- Publication Date
- 2021/10/01
- Publicized
- 2021/07/09
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.2021EDP7060
- Type of Manuscript
- PAPER
- Category
- Image Recognition, Computer Vision
Authors
Ying KANG
Defense Innovation Institute,Chinese People's Liberation Army
Cong LIU
National University of Defense Technology
Ning WANG
National University of Defense Technology
Dianxi SHI
Defense Innovation Institute,Tianjin Artificial Intelligence Innovation Center
Ning ZHOU
Chinese People's Liberation Army
Mengmeng LI
Defense Innovation Institute,Tianjin Artificial Intelligence Innovation Center
Yunlong WU
Defense Innovation Institute,Tianjin Artificial Intelligence Innovation Center
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Ying KANG, Cong LIU, Ning WANG, Dianxi SHI, Ning ZHOU, Mengmeng LI, Yunlong WU, "Siamese Visual Tracking with Dual-Pipeline Correlated Fusion Network" in IEICE TRANSACTIONS on Information,
vol. E104-D, no. 10, pp. 1702-1711, October 2021, doi: 10.1587/transinf.2021EDP7060.
Abstract: Siamese visual tracking, viewed as a problem of max-similarity matching to the target template, has absorbed increasing attention in computer vision. However, it is a challenge for current Siamese trackers that the demands of balance between accuracy in real-time tracking and robustness in long-time tracking are hard to meet. This work proposes a new Siamese based tracker with a dual-pipeline correlated fusion network (named as ADF-SiamRPN), which consists of one initial template for robust correlation, and the other transient template with the ability of adaptive feature optimal selection for accurate correlation. By the promotion from the learnable correlation-response fusion network afterwards, we are in pursuit of the synthetical improvement of tracking performance. To compare the performance of ADF-SiamRPN with state-of-the-art trackers, we conduct lots of experiments on benchmarks like OTB100, UAV123, VOT2016, VOT2018, GOT-10k, LaSOT and TrackingNet. The experimental results of tracking demonstrate that ADF-SiamRPN outperforms all the compared trackers and achieves the best balance between accuracy and robustness.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDP7060/_p
Copy
@ARTICLE{e104-d_10_1702,
author={Ying KANG, Cong LIU, Ning WANG, Dianxi SHI, Ning ZHOU, Mengmeng LI, Yunlong WU, },
journal={IEICE TRANSACTIONS on Information},
title={Siamese Visual Tracking with Dual-Pipeline Correlated Fusion Network},
year={2021},
volume={E104-D},
number={10},
pages={1702-1711},
abstract={Siamese visual tracking, viewed as a problem of max-similarity matching to the target template, has absorbed increasing attention in computer vision. However, it is a challenge for current Siamese trackers that the demands of balance between accuracy in real-time tracking and robustness in long-time tracking are hard to meet. This work proposes a new Siamese based tracker with a dual-pipeline correlated fusion network (named as ADF-SiamRPN), which consists of one initial template for robust correlation, and the other transient template with the ability of adaptive feature optimal selection for accurate correlation. By the promotion from the learnable correlation-response fusion network afterwards, we are in pursuit of the synthetical improvement of tracking performance. To compare the performance of ADF-SiamRPN with state-of-the-art trackers, we conduct lots of experiments on benchmarks like OTB100, UAV123, VOT2016, VOT2018, GOT-10k, LaSOT and TrackingNet. The experimental results of tracking demonstrate that ADF-SiamRPN outperforms all the compared trackers and achieves the best balance between accuracy and robustness.},
keywords={},
doi={10.1587/transinf.2021EDP7060},
ISSN={1745-1361},
month={October},}
Copy
TY - JOUR
TI - Siamese Visual Tracking with Dual-Pipeline Correlated Fusion Network
T2 - IEICE TRANSACTIONS on Information
SP - 1702
EP - 1711
AU - Ying KANG
AU - Cong LIU
AU - Ning WANG
AU - Dianxi SHI
AU - Ning ZHOU
AU - Mengmeng LI
AU - Yunlong WU
PY - 2021
DO - 10.1587/transinf.2021EDP7060
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E104-D
IS - 10
JA - IEICE TRANSACTIONS on Information
Y1 - October 2021
AB - Siamese visual tracking, viewed as a problem of max-similarity matching to the target template, has absorbed increasing attention in computer vision. However, it is a challenge for current Siamese trackers that the demands of balance between accuracy in real-time tracking and robustness in long-time tracking are hard to meet. This work proposes a new Siamese based tracker with a dual-pipeline correlated fusion network (named as ADF-SiamRPN), which consists of one initial template for robust correlation, and the other transient template with the ability of adaptive feature optimal selection for accurate correlation. By the promotion from the learnable correlation-response fusion network afterwards, we are in pursuit of the synthetical improvement of tracking performance. To compare the performance of ADF-SiamRPN with state-of-the-art trackers, we conduct lots of experiments on benchmarks like OTB100, UAV123, VOT2016, VOT2018, GOT-10k, LaSOT and TrackingNet. The experimental results of tracking demonstrate that ADF-SiamRPN outperforms all the compared trackers and achieves the best balance between accuracy and robustness.
ER -
English
