One-class support vector machine-assisted robust tracking
Artikel i vetenskaplig tidskrift, 2013

Recently, tracking is regarded as a binary classification problem by discriminative tracking methods. However, such binary classification may not fully handle the outliers, which may cause drifting. We argue that tracking may be regarded as one-class problem, which avoids gathering limited negative samples for background description. Inspired by the fact the positive feature space generated by one-class support vector machine (SVM) is bounded by a closed hyper sphere, we propose a tracking method utilizing one-class SVMs that adopt histograms of oriented gradient and 2bit binary patterns as features. Thus, it is called the one-class SVM tracker (OCST). Simultaneously, an efficient initialization and online updating scheme is proposed. Extensive experimental results prove that OCST outperforms some state-of-the-art discriminative tracking methods that tackle the problem using binary classifiers on providing accurate tracking and alleviating serious drifting.

support vector machine

visual object tracking

multiple instant learning

detection and tracking

Författare

Keren Fu

Ministry of Education China

Shanghai Jiao Tong University

Chen Gong

Ministry of Education China

Shanghai Jiao Tong University

Yu Qiao

Ministry of Education China

Shanghai Jiao Tong University

Jie Yang

Ministry of Education China

Shanghai Jiao Tong University

Irene Yu-Hua Gu

Chalmers, Signaler och system, Signalbehandling och medicinsk teknik

Journal of Electronic Imaging

1017-9909 (ISSN) 1560229x (eISSN)

Vol. 22 2 11- 023002

Styrkeområden

Informations- och kommunikationsteknik

Transport

Ämneskategorier

Systemvetenskap

Signalbehandling

Datorseende och robotik (autonoma system)

DOI

10.1117/1.JEI.22.2.023002

Mer information

Senast uppdaterat

2018-03-07