アラカワ カオル   Arakawa Kaoru
  荒川 薫
   所属   明治大学  総合数理学部
   職種   専任教授
言語種別 英語
発行・発表の年月 2018/11
形態種別 国際会議議事録
査読 査読有り
標題 Kernel Correlation Filter Tracker via Adaptive Model
執筆形態 共著(筆頭者以外)
掲載誌名 Proc. ISPACS 2018
掲載区分国外
出版社・発行元 IEEE
巻・号・頁 pp.205-209
総ページ数 5
著者・共著者 Tang Zhaoqian, Kaoru Arakawa
概要 In this paper, we propose a robust object tracking algorithm using an adaptive model and robust state recognition in kernel
correlation filter tracker. In the first stage of the proposed algorithm, we apply a scale pool technique to deal with scale
variation in object tracking. Then, the detection response from kernel correlation filter tracker is combined with grayscale
histogram similarity to estimate the state of the object. Furthermore, the classifier model is updated with adjustable learning
rate, thereby enabling the tracker to be robust to the change of the state of the object. Experimental results demonstrate that
the proposed tracker realizes outstanding performance on a challenging benchmark (OTB).
ISBN 978-1-5386-5770-6