Please use this identifier to cite or link to this item:
Scopus Web of Science® Altmetric
Type: Conference paper
Title: Part-based visual tracking with online latent structural learning
Author: Yao, R.
Shi, Q.
Shen, C.
Zhang, Y.
Van Den Hengel, A.
Citation: Proceedings, 2013 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2013, 23-28 June 2013, Portland, Oregon, USA: pp. 2363-2370
Publisher: IEEE
Publisher Place: United States of America
Issue Date: 2013
Series/Report no.: IEEE Conference on Computer Vision and Pattern Recognition
ISBN: 9780769549897
ISSN: 1063-6919
Conference Name: IEEE Conference on Computer Vision and Pattern Recognition (26th : 2013 : Portland, Oregon)
Statement of
Rui Yao, Qinfeng Shi, Chunhua Shen, Yanning Zhang, Anton van den Hengel
Abstract: Despite many advances made in the area, deformable targets and partial occlusions continue to represent key problems in visual tracking. Structured learning has shown good results when applied to tracking whole targets, but applying this approach to a part-based target model is complicated by the need to model the relationships between parts, and to avoid lengthy initialisation processes. We thus propose a method which models the unknown parts using latent variables. In doing so we extend the online algorithm pegasos to the structured prediction case (i.e., predicting the location of the bounding boxes) with latent part variables. To better estimate the parts, and to avoid over-fitting caused by the extra model complexity/capacity introduced by the parts, we propose a two-stage training process, based on the primal rather than the dual form. We then show that the method outperforms the state-of-the-art (linear and non-linear kernel) trackers.
Keywords: Online Latent Structural Learning; part-based tracking; visual tracking
Rights: ©IEEE
RMID: 0020133295
DOI: 10.1109/CVPR.2013.306
Grant ID:
Description (link):
Appears in Collections:Computer Science publications

Files in This Item:
File Description SizeFormat 
hdl_77413.pdfAccepted version2.22 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.