Skip to main navigation Skip to search Skip to main content

REVT: Robust and Efficient Visual Tracking by Region-Convolutional Regression Network

  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper proposes a novel approach, namely REVT, for visual tracking based on a region convolutional regression network. REVT runs according to a coarse to fine scheme. It first builds an on-line update deep network to roughly select a candidate region in a fast way. It then refines the result by exquisitely searching the target within the candidate region by a deep regression network, which is trained off-line to account for more diverse intra-class appearance changes. REVT thus integrates the advantages of the two types of deep models, and demonstrates a good trade-off between accuracy and efficiency. We perform extensive experiments on the OTB-2013 and OTB-2015 benchmarks, and REVT reports competitive performance at a speed of 19 fps, proving its competency.

Original languageEnglish
Title of host publicationMultiMedia Modeling - 24th International Conference, MMM 2018, Proceedings
EditorsKlaus Schoeffmann, Moncef Gabbouj, Noel E. O'Connor, Ahmed Elgammal, Thanarat H. Chalidabhongse, Supavadee Aramvith, Chong Wah Ngo, Yo-Sung Ho
PublisherSpringer Verlag
Pages440-452
Number of pages13
ISBN (Print)9783319736020
DOIs
StatePublished - 2018
Event24th International Conference on MultiMedia Modeling, MMM 2018 - Bangkok, Thailand
Duration: 5 Feb 20187 Feb 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10704 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th International Conference on MultiMedia Modeling, MMM 2018
Country/TerritoryThailand
CityBangkok
Period5/02/187/02/18

Keywords

  • Coarse-to-fine
  • Deep regression network
  • Visual tracking

Fingerprint

Dive into the research topics of 'REVT: Robust and Efficient Visual Tracking by Region-Convolutional Regression Network'. Together they form a unique fingerprint.

Cite this