Please use this identifier to cite or link to this item:
|Scopus||Web of Science®||Altmetric|
|Title:||Visual tracking with spatio-temporal Dempster-Shafer information fusion|
Van Den Hengel, A.
|Citation:||IEEE Transactions on Image Processing, 2013; 22(8):3028-3040|
|Publisher:||IEEE-Inst Electrical Electronics Engineers Inc|
|Xi Li, Anthony Dick, Chunhua Shen, Zhongfei Zhang, Anton van den Hengel, Hanzi Wang|
|Abstract:||A key problem in visual tracking is how to effectively combine spatio-temporal visual information from throughout a video to accurately estimate the state of an object. We address this problem by incorporating Dempster-Shafer information fusion into the tracking approach. To implement this fusion task, the entire image sequence is partitioned into spatially and temporally adjacent subsequences. A support vector machine (SVM) classifier is trained for object=non-object classification on each of these subsequences, the outputs of which act as separate data sources. To combine the discriminative information from these classifiers, we further present a spatio-temporal weighted Dempster-Shafer (STWDS) scheme. Moreover, temporally adjacent sources are likely to share discriminative information on object/non-object classification. In order to use such information, an adaptive SVM learning scheme is designed to transfer discriminative information across sources. Finally, the corresponding Dempster-Shafer belief function of the STWDS scheme is embedded into a Bayesian tracking model. Experimental results on challenging videos demonstrate the effectiveness and robustness of the proposed tracking approach.|
|Keywords:||Dempster- Shafer information fusion; Visual tracking; adaptive SVM learning; appearance model; multi-source discriminative learning|
|Rights:||Copyright (c) 2013 IEEE.|
|Appears in Collections:||Computer Science publications|
Files in This Item:
|hdl_77448.pdf||Accepted version||2.16 MB||Adobe PDF||View/Open|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.