Please use this identifier to cite or link to this item:
Type: Conference paper
Title: Matting-driven online learning of Hough forests for object tracking
Author: Qin, T.
Zhong, B.
Chin, T.J.
Wang, H.
Citation: Proceedings of the 21st International Conference on Pattern Recognition, held in Tsukuba, Japan, 11-15 November, 2012: pp.2488-2491
Publisher: IEEE
Publisher Place: USA
Issue Date: 2012
ISBN: 9781467322164
ISSN: 1051-4651
Conference Name: International Conference on Pattern Recognition (21st : 2012 : Tsukuba, Japan)
Statement of
Tao Qin, Bineng Zhong, Tat-Jun Chin and Hanzi Wang
Abstract: Accurate segmentation provides a useful contour constraint to alleviate drifting during online learning for tracking. Towards this end, we present a closed loop method for object tracking that links Hough forests and alpha matting via an effective back-projection scheme for patches. A novel hybrid-Houghforests-based method first estimates object location. Given the object location, the trimap of matting is then automatically generated from the patches backprojected from the Hough forests. Subsequently, an accurate contour of the object can be obtained based on a robust matting technique. Based on such an accurate contour, an update strategy is utilized to obtain reliably labeled samples to update the Hough forests to decrease the risk of model drift. Extensive comparisons on challenging sequences demonstrate the robustness and effectiveness of the proposed method.
Keywords: Object tracking; robustness; target tracking; vectors; vegetation; video sequences
Rights: © 2012 ICPR
RMID: 0020126326
Published version:
Appears in Collections:Computer Science publications

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.