Object tracking with convolutional neural networks and kernelized correlation filters

Dongxuan Li; Wenjie Chen

doi:10.1109/CCDC.2017.7978672

Object tracking with convolutional neural networks and kernelized correlation filters

Dongxuan Li, Wenjie Chen

School of Automation

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

Convolutional neural networks are widely used in object recognition and detection. In recent years, some researchers attempt to apply deep neural networks to visual object tracking. However, deep networks are extremely time-consuming and object tracking is not a classification problem essentially. In this paper, we present an online tracking framework which combines shallow convolutional neural networks with kernelized correlation filters(KCF). Different from offline training, our method successfully gets the convolution kernels by K-means clustering algorithm. Experimental results based on a representative visual tracker benchmark dataset show that the proposed method achieves excellent performance.

Original language	English
Title of host publication	Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1039-1044
Number of pages	6
ISBN (Electronic)	9781509046560
DOIs	https://doi.org/10.1109/CCDC.2017.7978672
Publication status	Published - 12 Jul 2017
Event	29th Chinese Control and Decision Conference, CCDC 2017 - Chongqing, China Duration: 28 May 2017 → 30 May 2017

Publication series

Name	Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017

Conference

Conference	29th Chinese Control and Decision Conference, CCDC 2017
Country/Territory	China
City	Chongqing
Period	28/05/17 → 30/05/17

Keywords

Convolutional Neural Networks
Kernelized Correlation Filters
Object Tracking

Access to Document

10.1109/CCDC.2017.7978672

Cite this

Li, D., & Chen, W. (2017). Object tracking with convolutional neural networks and kernelized correlation filters. In Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017 (pp. 1039-1044). Article 7978672 (Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CCDC.2017.7978672

@inproceedings{a61a3add3a2c45c68d5268bb89b416e2,

title = "Object tracking with convolutional neural networks and kernelized correlation filters",

abstract = "Convolutional neural networks are widely used in object recognition and detection. In recent years, some researchers attempt to apply deep neural networks to visual object tracking. However, deep networks are extremely time-consuming and object tracking is not a classification problem essentially. In this paper, we present an online tracking framework which combines shallow convolutional neural networks with kernelized correlation filters(KCF). Different from offline training, our method successfully gets the convolution kernels by K-means clustering algorithm. Experimental results based on a representative visual tracker benchmark dataset show that the proposed method achieves excellent performance.",

keywords = "Convolutional Neural Networks, Kernelized Correlation Filters, Object Tracking",

author = "Dongxuan Li and Wenjie Chen",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; 29th Chinese Control and Decision Conference, CCDC 2017 ; Conference date: 28-05-2017 Through 30-05-2017",

year = "2017",

month = jul,

day = "12",

doi = "10.1109/CCDC.2017.7978672",

language = "English",

series = "Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1039--1044",

booktitle = "Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017",

address = "United States",

}

Li, D & Chen, W 2017, Object tracking with convolutional neural networks and kernelized correlation filters. in Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017., 7978672, Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017, Institute of Electrical and Electronics Engineers Inc., pp. 1039-1044, 29th Chinese Control and Decision Conference, CCDC 2017, Chongqing, China, 28/05/17. https://doi.org/10.1109/CCDC.2017.7978672

Object tracking with convolutional neural networks and kernelized correlation filters. / Li, Dongxuan; Chen, Wenjie.
Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 1039-1044 7978672 (Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Object tracking with convolutional neural networks and kernelized correlation filters

AU - Li, Dongxuan

AU - Chen, Wenjie

PY - 2017/7/12

Y1 - 2017/7/12

N2 - Convolutional neural networks are widely used in object recognition and detection. In recent years, some researchers attempt to apply deep neural networks to visual object tracking. However, deep networks are extremely time-consuming and object tracking is not a classification problem essentially. In this paper, we present an online tracking framework which combines shallow convolutional neural networks with kernelized correlation filters(KCF). Different from offline training, our method successfully gets the convolution kernels by K-means clustering algorithm. Experimental results based on a representative visual tracker benchmark dataset show that the proposed method achieves excellent performance.

AB - Convolutional neural networks are widely used in object recognition and detection. In recent years, some researchers attempt to apply deep neural networks to visual object tracking. However, deep networks are extremely time-consuming and object tracking is not a classification problem essentially. In this paper, we present an online tracking framework which combines shallow convolutional neural networks with kernelized correlation filters(KCF). Different from offline training, our method successfully gets the convolution kernels by K-means clustering algorithm. Experimental results based on a representative visual tracker benchmark dataset show that the proposed method achieves excellent performance.

KW - Convolutional Neural Networks

KW - Kernelized Correlation Filters

KW - Object Tracking

UR - http://www.scopus.com/inward/record.url?scp=85028027625&partnerID=8YFLogxK

U2 - 10.1109/CCDC.2017.7978672

DO - 10.1109/CCDC.2017.7978672

M3 - Conference contribution

AN - SCOPUS:85028027625

T3 - Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017

SP - 1039

EP - 1044

BT - Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 29th Chinese Control and Decision Conference, CCDC 2017

Y2 - 28 May 2017 through 30 May 2017

ER -

Li D, Chen W. Object tracking with convolutional neural networks and kernelized correlation filters. In Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 1039-1044. 7978672. (Proceedings of the 29th Chinese Control and Decision Conference, CCDC 2017). doi: 10.1109/CCDC.2017.7978672

Object tracking with convolutional neural networks and kernelized correlation filters

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this