Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning

Yifan Jiang, Hyunhak Shin, Hanseok Ko

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

In this paper, we propose a precise regression approach for correcting imprecise bounding box using deep reinforcement learning. Object tracking task essentially builds trajectory of a moving object based on detection and tracking algorithms and its current state is indicated by having the object encapsulated with a bounding box corresponding to its position and size. However due to the imperfect detection and tracking algorithms operating in complex scene, it is difficult to obtain the precise bounding box as errors frequently occur producing oversized, partial, and false bounding box, respectively. To correct the error, we train an intelligent agent that move the bounding box to the right position and scale it to its correct size matching to that of the true target. The agent is trained by deep Q-Iearning and evaluated on several state-of-the-art multiple object tracking approaches. The experimental results demonstrate that our proposed framework can effectively eliminate the object tracking bounding box error and its robustness is verified by realizing improved tracking performance in complex scene.

Original languageEnglish
Title of host publication2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1643-1647
Number of pages5
Volume2018-April
ISBN (Print)9781538646588
DOIs
Publication statusPublished - 2018 Sep 10
Event2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Calgary, Canada
Duration: 2018 Apr 152018 Apr 20

Other

Other2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
CountryCanada
CityCalgary
Period18/4/1518/4/20

Fingerprint

Reinforcement learning
Intelligent agents
Trajectories

Keywords

  • Bounding box
  • Object tracking
  • Regression
  • Reinforcement learning

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Jiang, Y., Shin, H., & Ko, H. (2018). Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings (Vol. 2018-April, pp. 1643-1647). [8462063] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2018.8462063

Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning. / Jiang, Yifan; Shin, Hyunhak; Ko, Hanseok.

2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. p. 1643-1647 8462063.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Jiang, Y, Shin, H & Ko, H 2018, Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning. in 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. vol. 2018-April, 8462063, Institute of Electrical and Electronics Engineers Inc., pp. 1643-1647, 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018, Calgary, Canada, 18/4/15. https://doi.org/10.1109/ICASSP.2018.8462063
Jiang Y, Shin H, Ko H. Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April. Institute of Electrical and Electronics Engineers Inc. 2018. p. 1643-1647. 8462063 https://doi.org/10.1109/ICASSP.2018.8462063
Jiang, Yifan ; Shin, Hyunhak ; Ko, Hanseok. / Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning. 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers Inc., 2018. pp. 1643-1647
@inproceedings{248c8da3f6864ad8b1cde3a9976801ae,
title = "Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning",
abstract = "In this paper, we propose a precise regression approach for correcting imprecise bounding box using deep reinforcement learning. Object tracking task essentially builds trajectory of a moving object based on detection and tracking algorithms and its current state is indicated by having the object encapsulated with a bounding box corresponding to its position and size. However due to the imperfect detection and tracking algorithms operating in complex scene, it is difficult to obtain the precise bounding box as errors frequently occur producing oversized, partial, and false bounding box, respectively. To correct the error, we train an intelligent agent that move the bounding box to the right position and scale it to its correct size matching to that of the true target. The agent is trained by deep Q-Iearning and evaluated on several state-of-the-art multiple object tracking approaches. The experimental results demonstrate that our proposed framework can effectively eliminate the object tracking bounding box error and its robustness is verified by realizing improved tracking performance in complex scene.",
keywords = "Bounding box, Object tracking, Regression, Reinforcement learning",
author = "Yifan Jiang and Hyunhak Shin and Hanseok Ko",
year = "2018",
month = "9",
day = "10",
doi = "10.1109/ICASSP.2018.8462063",
language = "English",
isbn = "9781538646588",
volume = "2018-April",
pages = "1643--1647",
booktitle = "2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Precise Regression for Bounding Box Correction for Improved Tracking Based on Deep Reinforcement Learning

AU - Jiang, Yifan

AU - Shin, Hyunhak

AU - Ko, Hanseok

PY - 2018/9/10

Y1 - 2018/9/10

N2 - In this paper, we propose a precise regression approach for correcting imprecise bounding box using deep reinforcement learning. Object tracking task essentially builds trajectory of a moving object based on detection and tracking algorithms and its current state is indicated by having the object encapsulated with a bounding box corresponding to its position and size. However due to the imperfect detection and tracking algorithms operating in complex scene, it is difficult to obtain the precise bounding box as errors frequently occur producing oversized, partial, and false bounding box, respectively. To correct the error, we train an intelligent agent that move the bounding box to the right position and scale it to its correct size matching to that of the true target. The agent is trained by deep Q-Iearning and evaluated on several state-of-the-art multiple object tracking approaches. The experimental results demonstrate that our proposed framework can effectively eliminate the object tracking bounding box error and its robustness is verified by realizing improved tracking performance in complex scene.

AB - In this paper, we propose a precise regression approach for correcting imprecise bounding box using deep reinforcement learning. Object tracking task essentially builds trajectory of a moving object based on detection and tracking algorithms and its current state is indicated by having the object encapsulated with a bounding box corresponding to its position and size. However due to the imperfect detection and tracking algorithms operating in complex scene, it is difficult to obtain the precise bounding box as errors frequently occur producing oversized, partial, and false bounding box, respectively. To correct the error, we train an intelligent agent that move the bounding box to the right position and scale it to its correct size matching to that of the true target. The agent is trained by deep Q-Iearning and evaluated on several state-of-the-art multiple object tracking approaches. The experimental results demonstrate that our proposed framework can effectively eliminate the object tracking bounding box error and its robustness is verified by realizing improved tracking performance in complex scene.

KW - Bounding box

KW - Object tracking

KW - Regression

KW - Reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85054227691&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054227691&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2018.8462063

DO - 10.1109/ICASSP.2018.8462063

M3 - Conference contribution

AN - SCOPUS:85054227691

SN - 9781538646588

VL - 2018-April

SP - 1643

EP - 1647

BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -