Summary
International Technical Conference on Circuits/Systems, Computers and Communications
2016
Session Number:T2-5
Session:
Number:4874
Object Tracking based on Path Similarity of Boosted Decision Trees
Koichi Mitsunari, Jaehoon Yu, Yoshinori Takeuchi, Masaharu Imai ,
pp.563-566
Publication Date:2016/7/10
Online ISSN:2188-5079
DOI:10.34385/proc.61.4874
PDF download (1.1MB)
Summary:
For general object recognition, detection and tracking are important and complementary components but require different types of features. Therefore, it is difficult to fuse two components. To address the issue, we present a novel tracking method using path similarities between binary decision trees, which are byproducts obtained from detection process. In the experimental result, the proposed method achieved comparable tracking capability to a conventional tracking method using on-line boosting without any extra computation.