Hedayati, M., Cree, M. J., & Scott, J. B. (2017). Effect of contextual information on object tracking. Presented at the 2017 International Conference on Image and Vision Computing New Zealand (IVCNZ), Washington, DC, USA: IEEE. https://doi.org/10.1109/IVCNZ.2017.8402488
Permanent Research Commons link: https://hdl.handle.net/10289/12002
Local object information, such as the appearance and motion features of the object, are useful for object tracking in videos provided the object is not occluded by other elements in the scene. During occlusion, however, the local object information in the video frame does not properly represent the true properties of the object, which leads to tracking failure. We propose a framework that combines multiple cues including the local object information, the background characteristics and group motion dynamics to improve object tracking in challenging cluttered environments. The performance of the proposed tracking model is compared with the kernelised correlation filter (KCF) tracker. In the tested video sequences the proposed tracking model correctly tracked objects even when the KCF tracker failed because of occlusion and background noise.
© 2018 IEEE.This is an author’s accepted version. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.