Institutional Repository of Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
Feature Distilled Tracking | |
Zhu Guibo1![]() ![]() ![]() ![]() | |
Source Publication | IEEE Transaction on Cybernetics
![]() |
2017-12 | |
Issue | 0Pages:0 |
Abstract |
Feature extraction and representation is one of the most important components for fast, accurate, and robust visual tracking. Very deep convolutional neural networks (CNNs) provide effective tools for feature extraction with good generalization ability. However, extracting features using very deep CNN models needs high performance hardware due to its large computation complexity, which prohibits its extensions in real-time applications. To alleviate this problem, we aim at obtaining small and fast-to-execute shallow models based on model compression for visual tracking. Specifically, we propose a small feature distilled network (FDN) for tracking by imitating the intermediate representations of a much deeper network. The FDN extracts rich visual features with higher speed than the original deeper network. To further speed-up, we introduce a shift-and-stitch method to reduce the arithmetic operations, while preserving the spatial resolution of the distilled feature maps unchanged. Finally, a scale adaptive discriminative correlation filter is learned on the distilled feature for visual tracking to handle scale variation of the target. Comprehensive experimental results on object tracking benchmark datasets show that the proposed approach achieves 5x speed-up with competitive performance to the state-of-the-art deep trackers.
|
Keyword | Correlation Filter Model Compression Visual Tracking |
Subject Area | Computer Science, Artificial Intelligence, Cybernetics |
DOI | 10.1109/TCYB.2017.2776977 |
Indexed By | SCI |
Language | 英语 |
Funding Organization | National Natural Science Foundation of China under Grant 61702510, Grant 61773375, Grant 61370036, Grant 61772277, and Grant 61772527. |
WOS ID | WOS:3 |
Citation statistics | |
Document Type | 期刊论文 |
Identifier | http://ir.ia.ac.cn/handle/173211/22062 |
Collection | 模式识别国家重点实验室_图像与视频分析 |
Corresponding Author | Jinqiao Wang |
Affiliation | 1.Institute of Automation, Chinese Academy of Sciences 2.University of Chinese Academy of Sciences 3.Nanjing Audit University 4.Indiana University School of Medicine |
First Author Affilication | Institute of Automation, Chinese Academy of Sciences |
Corresponding Author Affilication | Institute of Automation, Chinese Academy of Sciences |
Recommended Citation GB/T 7714 | Zhu Guibo,Jinqiao Wang,Peisong Wang,et al. Feature Distilled Tracking[J]. IEEE Transaction on Cybernetics,2017(0):0. |
APA | Zhu Guibo,Jinqiao Wang,Peisong Wang,Yi Wu,&Hanqing Lu.(2017).Feature Distilled Tracking.IEEE Transaction on Cybernetics(0),0. |
MLA | Zhu Guibo,et al."Feature Distilled Tracking".IEEE Transaction on Cybernetics .0(2017):0. |
Files in This Item: | Download All | |||||
File Name/Size | DocType | Version | Access | License | ||
feature distilled tr(2450KB) | 期刊论文 | 作者接受稿 | 开放获取 | CC BY-NC-SA | View Download |
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment