Online Boosted Tracking with Discriminative Feature Selection and Scale Adaptation

Hefeng Wu,Guanbin Li,Zhuo Su,Xiaonan Luo
DOI: https://doi.org/10.1109/icip.2012.6466880
2012-01-01
Abstract:We track the object by separating it from the surrounding with an ensemble of boosted classifiers, which are trained in a discriminative feature space that is determined on the fly. Contour refinement and weight thresholding techniques are used to select good examples for training. While tracking, location calibration and scale adaptation are used to improve the tracker's performance. We update the ensemble of weak classifiers online to adapt to appearance changes, and use the positive occupancy ratio to detect occlusion. A center-surround discrepancy measure is presented to evaluate the discriminative power of the current feature space and to invoke re-initialization of feature selection and classifier training if necessary. Experiments on challenging video sequences demonstrate the effectiveness of the proposed approach.
What problem does this paper attempt to address?