8 September 2016 Object tracking with hierarchical multiview learning
Author Affiliations +
Building a robust appearance model is useful to improve tracking performance. We propose a hierarchical multiview learning framework to construct the appearance model, which has two layers for tracking. On the top layer, two different views of features, grayscale value and histogram of oriented gradients, are adopted for representation under the cotraining framework. On the bottom layer, for each view of each feature, three different random subspaces are generated to represent the appearance from multiple views. For each random view submodel, the least squares support vector machine is employed to improve the discriminability for concrete and efficient realization. These two layers are combined to construct the final appearance model for tracking. The proposed hierarchical model assembles two types of multiview learning strategies, in which the appearance can be described more accurately and robustly. Experimental results in the benchmark dataset demonstrate that the proposed method can achieve better performance than many existing state-of-the-art algorithms.
© 2016 SPIE and IS&T
Jun Yang, Jun Yang, Shunli Zhang, Shunli Zhang, Li Zhang, Li Zhang, } "Object tracking with hierarchical multiview learning," Journal of Electronic Imaging 25(5), 053006 (8 September 2016). https://doi.org/10.1117/1.JEI.25.5.053006 . Submission:


Model recommendation for pedestrian detection
Proceedings of SPIE (August 28 2016)
Emotion detection model of Filipino music
Proceedings of SPIE (February 07 2017)
Feature distributions and probability of identification
Proceedings of SPIE (August 12 1999)
Font adaptation of an HMM-based OCR system
Proceedings of SPIE (January 18 2010)

Back to Top