Kernel-based Online Object Tracking via Gaussian Mixture Model Learning
Object tracking has attracted increasing attention and requires difficult object appearance recognition and learning.This paper proposes a novel object tracking by matching corresponding SURF-based keypoints.A dynamic 2D scalerotation space is constructed for each object keypoint to strengthen its variation and distinctiveness.During matching,we assign each keypoint a kernel weight employing Gaussian mixture model,to make sure those keypoints with higher repeatability and reliability are used.Meanwhile,improved weighted RANSAC is applied to estimate motion parameters.Finally,on-line learning is performed on SURF feature,2D scalerotation space and Gaussian mixture model once tracking is successful.Experimental results using both public and private image sequences validate the robustness and accuracy of the proposed method under complex scene changes.
objec tracking kernel function Gaussian mixuture model feature matching online learning
Quan Miao Yanfeng Gu
National Computer Network Emergency Response Technical Team/Coordination Center of China Beijing, Ch School of Electronics and Information Engineering Harbin Institute of Technology Harbin, China
国际会议
哈尔滨
英文
522-525
2016-07-21（万方平台首次上网日期，不代表论文的发表时间）