Skip navigation
Please use this identifier to cite or link to this item:
Title: Extraction of long-duration moving object trajectories from curtailed tracks
Authors: Ahmed S.A.
Dogra D.P.
Kar S.
Pratim Roy, Partha
Chaudhuri B.B.
Raman B.
Kankanhalli M.
Published in: Proceedings of Advances in Intelligent Systems and Computing
Abstract: Object tracking remains one of the critical challenges in visual surveillance. It is difficult to track each moving object in a crowded scene. This paper proposed a new approach to track moving objects for longer duration. First, key points are tracked for short duration using state-of-the-art feature tracker. Next, the features are grouped and linked in spatiotemporal domain. Finally, we create a single trajectory for each object or a group of similar objects. We have tested the method on publicly available video datasets where more than 100 people were moving randomly. The results reveal that the proposed method can be highly effective to extract long-duration trajectories from the curtailed tracklets obtained using short-duration feature tracker. © Springer Nature Singapore Pte Ltd. 2018.
Citation: Proceedings of Advances in Intelligent Systems and Computing, (2018), 315- 326
Issue Date: 2018
Publisher: Springer Verlag
Keywords: Trajectories
Critical challenges
Feature tracker
Moving object trajectories
Object Tracking
Short durations
Spatio-temporal domains
State of the art
Visual surveillance
Computer vision
ISBN: 9.78981E+12
ISSN: 21945357
Author Scopus IDs: 57190343458
Author Affiliations: Ahmed, S.A., Haldia Institute of Technology, Haldia, India, National Institute of Technology Durgapur, Durgapur, India
Dogra, D.P., IIT Bhubaneswar, Bhubaneswar, 752050, India
Kar, S., National Institute of Technology Durgapur, Durgapur, India
Roy, P.P., Indian Institute of Technology Roorkee, Roorkee, India
Corresponding Author: Ahmed, S.A.; Haldia Institute of TechnologyIndia; email:
Appears in Collections:Conference Publications [CS]

Files in This Item:
There are no files associated with this item.
Show full item record

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