Correlation filter‐based visual tracking via adaptive weighted CNN features fusion. Issue 8 (1st August 2018)

• Record Type: Journal Article
• Authors: Hao, Zhaohui; Liu, Guixi; Zhang, Haoyang
• Is Part Of: IET image processing. Volume 12:Issue 8(2018)
• Page Start: 1423
• Page End: 1431
• Publication Date: 2018-08-01
• Subjects: neural nets -- image fusion -- video signal processing -- image sequences -- object tracking -- computer vision -- image representation -- object detection -- feature extraction; adaptive weighted CNN features fusion -- visual object tracking -- important task -- computer vision -- novel visual tracking approach -- tracking task -- scale estimation -- translation estimation -- multiple adaptive correlation filters -- hierarchical convolutional neural networks -- target location -- multilevel features -- different CNN layers -- adaptive weighted algorithm -- correlation response maps -- one‐dimensional correlation filter -- HOG features -- scale variation -- 50 challenging benchmark video sequences; Image processing -- Periodicals; 621.36705
• DOI: 10.1049/iet-ipr.2017.0443
• Languages: English
• ISSNs: 1751-9659

Abstract

Abstract : Visual object tracking is an important and challenging task in computer vision. In this study, the authors propose a novel visual tracking approach by decomposing the tracking task into translation and scale estimation. In translation estimation, they employ multiple adaptive correlation filters with features of hierarchical convolutional neural networks (CNNs) to more accurately estimate the target location. To make full use of multi‐level features from different CNN layers, they propose an adaptive weighted algorithm to fuse correlation response maps. In scale estimation, a one‐dimensional correlation filter with histogram of oriented gradient (HOG) features is employed to estimate the scale variation. Extensive experimental results on 50 challenging benchmark video sequences demonstrate that the proposed algorithm outperforms state‐of‐the‐art algorithms.