Persistent object tracking in road panoramic videos

Authors:
Yuan Zhou;Zhong Zhou;Ke Chen;Wei Wu
Affiliations:
State Key Laboratory of Virtual Reality Technology and Systems, Beihang University School of Computer Science and Engineering, Beijing, P.R. China;State Key Laboratory of Virtual Reality Technology and Systems, Beihang University School of Computer Science and Engineering, Beijing, P.R. China;State Key Laboratory of Virtual Reality Technology and Systems, Beihang University School of Computer Science and Engineering, Beijing, P.R. China;State Key Laboratory of Virtual Reality Technology and Systems, Beihang University School of Computer Science and Engineering, Beijing, P.R. China
Venue:
PCM'12 Proceedings of the 13th Pacific-Rim conference on Advances in Multimedia Information Processing
Year:
2012

Citing 4
Cited 0

Incremental Learning for Robust Visual Tracking

International Journal of Computer Vision
A Probabilistic Approach to Integrating Multiple Cues in Visual Tracking

ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part II
3D Panoramic Reconstruction with an Uncalibrated System of Stereovision Using Evolutionary Algorithms

SITIS '08 Proceedings of the 2008 IEEE International Conference on Signal Image Technology and Internet Based Systems
Static Object Tracking in Road Panoramic Videos

ISM '10 Proceedings of the 2010 IEEE International Symposium on Multimedia

Quantified Score

Hi-index	0.00

Visualization

Abstract

Panorama has the full directional view of the scene and can provide an object vision persistently from its emerging to vanishing except occlusion. Though, traditional tracking algorithms are apt to fail since the object may change its appearance or even disappear occasionally during its display lifetime. A persistent object tracking algorithm for static objects in panoramic videos is proposed in this paper. It creates several auxiliary trackers to guide the tracking. Once the object is obscured or deformed, the auxiliary trackers are engaged in estimating the position of the main tracker. Even though the appearance of the object changes a lot, its position still could be estimated with help of the auxiliary trackers. Experiment results illustrated that this algorithm provides a real-time tracking effectively on signs in road panoramic videos. This algorithm is easy to perform and especially valuable for road sign labeling and management.