Title :
Object Tracking With Only Background Cues
Author :
Annan Li ; Shuicheng Yan
Author_Institution :
Inst. for Infocomm Res., Agency for Sci., Technol. & Res., Singapore, Singapore
Abstract :
Background cues mainly play a supplementary or accompanying role in most previous approaches for object tracking. If object tracking is treated as a binary target/background classification problem, then the similarity with the target and the difference from the background can be considered equally informative. This leads to an interesting question: is it possible to perform object tracking using background cues only? To answer this question, we propose an object tracking approach that utilizes background cues only. The extensive experimental results positively validate the possibility of performing object tracking using background cues only. The results revealed in this paper also provide a new reference in designing future object tracking methods.
Keywords :
image classification; object detection; object tracking; target tracking; background classification problem; background cues; binary target classification problem; object tracking; Color; Computational modeling; Histograms; Image color analysis; Object tracking; Principal component analysis; Target tracking; Background cues; object tracking;
Journal_Title :
Circuits and Systems for Video Technology, IEEE Transactions on
DOI :
10.1109/TCSVT.2014.2317888