Maximum likelihood FIR filter for visual object tracking

Jung Min Pak; Choon Ki Ahn; Yung Hak Mo; Myo Taeg Lim; Moon Kyou Song

doi:10.1016/j.neucom.2016.07.047

Maximum likelihood FIR filter for visual object tracking

Jung Min Pak, Choon Ki Ahn, Yung Hak Mo, Myo Taeg Lim, Moon Kyou Song

Research output: Contribution to journal › Article › peer-review

19 Citations (Scopus)

Abstract

Visual object trackers usually adopt filters, such as the Kalman filter (KF) and the particle filter (PF), in order to improve tracking accuracy by suppressing measurement noises. However, if the filters have infinite impulse response (IIR) structures, the visual trackers adopting them can exhibit degraded tracking performance when system models have parameter uncertainties or when the noise information is incorrect. To overcome this problem, in this paper, we propose a new finite impulse response (FIR) filter for visual object tracking (VOT). The proposed filter is derived by maximizing the likelihood function, and it is referred to as the maximum likelihood FIR filter (MLFIRF). We conducted extensive experiments to show that the MLFIRF provides superior and more reliable tracking results compared with the KF, PF, and H_∞ filter (HF) in VOT.

Original language	English
Pages (from-to)	543-553
Number of pages	11
Journal	Neurocomputing
Volume	216
DOIs	https://doi.org/10.1016/j.neucom.2016.07.047
Publication status	Published - 2016 Dec 5

Bibliographical note

Funding Information:
This work was supported in part by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (Grant No. NRF-2016R1D1A1B01016071 ), in part by Basic Science Research Program through the NRF funded by the Ministry of Science, ICT, and Future Planning (Grant No. NRF-2014R1A1A1006101 ), and in part by Basic Science Research Program through the NRF funded by the Ministry of Education (Grant No. NRF-2013R1A1A2060663 ).

Publisher Copyright:
© 2016 Elsevier B.V.

Keywords

Finite impulse response (FIR) filter
Maximum likelihood
Maximum likelihood FIR filter (MLFIRF)
Visual object tracking (VOT)

ASJC Scopus subject areas

Computer Science Applications
Cognitive Neuroscience
Artificial Intelligence

Access to Document

10.1016/j.neucom.2016.07.047

Cite this

@article{2f1b4667c782428496d29af9e3629eee,

title = "Maximum likelihood FIR filter for visual object tracking",

abstract = "Visual object trackers usually adopt filters, such as the Kalman filter (KF) and the particle filter (PF), in order to improve tracking accuracy by suppressing measurement noises. However, if the filters have infinite impulse response (IIR) structures, the visual trackers adopting them can exhibit degraded tracking performance when system models have parameter uncertainties or when the noise information is incorrect. To overcome this problem, in this paper, we propose a new finite impulse response (FIR) filter for visual object tracking (VOT). The proposed filter is derived by maximizing the likelihood function, and it is referred to as the maximum likelihood FIR filter (MLFIRF). We conducted extensive experiments to show that the MLFIRF provides superior and more reliable tracking results compared with the KF, PF, and H∞ filter (HF) in VOT.",

keywords = "Finite impulse response (FIR) filter, Maximum likelihood, Maximum likelihood FIR filter (MLFIRF), Visual object tracking (VOT)",

author = "{Min Pak}, Jung and {Ki Ahn}, Choon and {Hak Mo}, Yung and {Taeg Lim}, Myo and {Kyou Song}, Moon",

note = "Funding Information: This work was supported in part by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (Grant No. NRF-2016R1D1A1B01016071 ), in part by Basic Science Research Program through the NRF funded by the Ministry of Science, ICT, and Future Planning (Grant No. NRF-2014R1A1A1006101 ), and in part by Basic Science Research Program through the NRF funded by the Ministry of Education (Grant No. NRF-2013R1A1A2060663 ). Publisher Copyright: {\textcopyright} 2016 Elsevier B.V.",

year = "2016",

month = dec,

day = "5",

doi = "10.1016/j.neucom.2016.07.047",

language = "English",

volume = "216",

pages = "543--553",

journal = "Neurocomputing",

issn = "0925-2312",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - Maximum likelihood FIR filter for visual object tracking

AU - Min Pak, Jung

AU - Ki Ahn, Choon

AU - Hak Mo, Yung

AU - Taeg Lim, Myo

AU - Kyou Song, Moon

N1 - Funding Information: This work was supported in part by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (Grant No. NRF-2016R1D1A1B01016071 ), in part by Basic Science Research Program through the NRF funded by the Ministry of Science, ICT, and Future Planning (Grant No. NRF-2014R1A1A1006101 ), and in part by Basic Science Research Program through the NRF funded by the Ministry of Education (Grant No. NRF-2013R1A1A2060663 ). Publisher Copyright: © 2016 Elsevier B.V.

PY - 2016/12/5

Y1 - 2016/12/5

N2 - Visual object trackers usually adopt filters, such as the Kalman filter (KF) and the particle filter (PF), in order to improve tracking accuracy by suppressing measurement noises. However, if the filters have infinite impulse response (IIR) structures, the visual trackers adopting them can exhibit degraded tracking performance when system models have parameter uncertainties or when the noise information is incorrect. To overcome this problem, in this paper, we propose a new finite impulse response (FIR) filter for visual object tracking (VOT). The proposed filter is derived by maximizing the likelihood function, and it is referred to as the maximum likelihood FIR filter (MLFIRF). We conducted extensive experiments to show that the MLFIRF provides superior and more reliable tracking results compared with the KF, PF, and H∞ filter (HF) in VOT.

AB - Visual object trackers usually adopt filters, such as the Kalman filter (KF) and the particle filter (PF), in order to improve tracking accuracy by suppressing measurement noises. However, if the filters have infinite impulse response (IIR) structures, the visual trackers adopting them can exhibit degraded tracking performance when system models have parameter uncertainties or when the noise information is incorrect. To overcome this problem, in this paper, we propose a new finite impulse response (FIR) filter for visual object tracking (VOT). The proposed filter is derived by maximizing the likelihood function, and it is referred to as the maximum likelihood FIR filter (MLFIRF). We conducted extensive experiments to show that the MLFIRF provides superior and more reliable tracking results compared with the KF, PF, and H∞ filter (HF) in VOT.

KW - Finite impulse response (FIR) filter

KW - Maximum likelihood

KW - Maximum likelihood FIR filter (MLFIRF)

KW - Visual object tracking (VOT)

UR - http://www.scopus.com/inward/record.url?scp=84994121203&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2016.07.047

DO - 10.1016/j.neucom.2016.07.047

M3 - Article

AN - SCOPUS:84994121203

SN - 0925-2312

VL - 216

SP - 543

EP - 553

JO - Neurocomputing

JF - Neurocomputing

ER -

Maximum likelihood FIR filter for visual object tracking

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this