Getting robust observation for single object tracking: A statistical kernel-based approach

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

4 Citations (Scopus)

Abstract

Mean shift-based algorithms perform well when the tracked object is in the vicinity of the current location. This cause any fast moving object especially when there is no overlapping region between the frames fails to be tracked. The aim of our algorithm is to offer robust kernel-based observation as an input to a single object tracking. We integrate kernel-based method with feature detectors and apply statical decision making. The foundation of the algorithm is patch matching where Epanechnikov kernel-based histogram is used to find the best patch. The patch is built based on Shi and Tomasi [1] corner detector where a vector descriptor is built at each detected corner. The patches are built at every matched points and the similarity between two histograms are modelled by Gaussian distribution. Two set of histograms are built based on RGB and HSV colour space where Neyman-Pearson method decides the best colour model. Diamond search configuration is applied to smooth out the patch position by applying maximum likelihood method. The works by Comaniciu et al. [2] is used as performance comparison. The results show that our algorithm performs better as we have no failure yet lesser average accuracy in tracking fast moving object.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages351-359
Number of pages9
Volume6854 LNCS
EditionPART 1
DOIs
Publication statusPublished - 2011
Externally publishedYes
Event14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011 - Seville
Duration: 29 Aug 201131 Aug 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume6854 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011
CitySeville
Period29/8/1131/8/11

Fingerprint

Object Tracking
Patch
kernel
Histogram
Moving Objects
Color
Detectors
Detector
Gaussian distribution
Mean Shift
Maximum likelihood
Color Space
Maximum Likelihood Method
Performance Comparison
Diamonds
Strombus or kite or diamond
Decision making
Descriptors
Overlapping
Decision Making

Keywords

  • Maximum likelihood
  • Neyman-Pearson
  • Patch matching
  • Tracking observation

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Zulkifley, M. A., & Moran, B. (2011). Getting robust observation for single object tracking: A statistical kernel-based approach. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 1 ed., Vol. 6854 LNCS, pp. 351-359). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6854 LNCS, No. PART 1). https://doi.org/10.1007/978-3-642-23672-3_43

Getting robust observation for single object tracking : A statistical kernel-based approach. / Zulkifley, Mohd Asyraf; Moran, Bill.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6854 LNCS PART 1. ed. 2011. p. 351-359 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6854 LNCS, No. PART 1).

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

Zulkifley, MA & Moran, B 2011, Getting robust observation for single object tracking: A statistical kernel-based approach. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 edn, vol. 6854 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 6854 LNCS, pp. 351-359, 14th International Conference on Computer Analysis of Images and Patterns, CAIP 2011, Seville, 29/8/11. https://doi.org/10.1007/978-3-642-23672-3_43
Zulkifley MA, Moran B. Getting robust observation for single object tracking: A statistical kernel-based approach. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 ed. Vol. 6854 LNCS. 2011. p. 351-359. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). https://doi.org/10.1007/978-3-642-23672-3_43
Zulkifley, Mohd Asyraf ; Moran, Bill. / Getting robust observation for single object tracking : A statistical kernel-based approach. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6854 LNCS PART 1. ed. 2011. pp. 351-359 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
@inproceedings{92f88c437f35490b90ab5039dde823b4,
title = "Getting robust observation for single object tracking: A statistical kernel-based approach",
abstract = "Mean shift-based algorithms perform well when the tracked object is in the vicinity of the current location. This cause any fast moving object especially when there is no overlapping region between the frames fails to be tracked. The aim of our algorithm is to offer robust kernel-based observation as an input to a single object tracking. We integrate kernel-based method with feature detectors and apply statical decision making. The foundation of the algorithm is patch matching where Epanechnikov kernel-based histogram is used to find the best patch. The patch is built based on Shi and Tomasi [1] corner detector where a vector descriptor is built at each detected corner. The patches are built at every matched points and the similarity between two histograms are modelled by Gaussian distribution. Two set of histograms are built based on RGB and HSV colour space where Neyman-Pearson method decides the best colour model. Diamond search configuration is applied to smooth out the patch position by applying maximum likelihood method. The works by Comaniciu et al. [2] is used as performance comparison. The results show that our algorithm performs better as we have no failure yet lesser average accuracy in tracking fast moving object.",
keywords = "Maximum likelihood, Neyman-Pearson, Patch matching, Tracking observation",
author = "Zulkifley, {Mohd Asyraf} and Bill Moran",
year = "2011",
doi = "10.1007/978-3-642-23672-3_43",
language = "English",
isbn = "9783642236716",
volume = "6854 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 1",
pages = "351--359",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 1",

}

TY - GEN

T1 - Getting robust observation for single object tracking

T2 - A statistical kernel-based approach

AU - Zulkifley, Mohd Asyraf

AU - Moran, Bill

PY - 2011

Y1 - 2011

N2 - Mean shift-based algorithms perform well when the tracked object is in the vicinity of the current location. This cause any fast moving object especially when there is no overlapping region between the frames fails to be tracked. The aim of our algorithm is to offer robust kernel-based observation as an input to a single object tracking. We integrate kernel-based method with feature detectors and apply statical decision making. The foundation of the algorithm is patch matching where Epanechnikov kernel-based histogram is used to find the best patch. The patch is built based on Shi and Tomasi [1] corner detector where a vector descriptor is built at each detected corner. The patches are built at every matched points and the similarity between two histograms are modelled by Gaussian distribution. Two set of histograms are built based on RGB and HSV colour space where Neyman-Pearson method decides the best colour model. Diamond search configuration is applied to smooth out the patch position by applying maximum likelihood method. The works by Comaniciu et al. [2] is used as performance comparison. The results show that our algorithm performs better as we have no failure yet lesser average accuracy in tracking fast moving object.

AB - Mean shift-based algorithms perform well when the tracked object is in the vicinity of the current location. This cause any fast moving object especially when there is no overlapping region between the frames fails to be tracked. The aim of our algorithm is to offer robust kernel-based observation as an input to a single object tracking. We integrate kernel-based method with feature detectors and apply statical decision making. The foundation of the algorithm is patch matching where Epanechnikov kernel-based histogram is used to find the best patch. The patch is built based on Shi and Tomasi [1] corner detector where a vector descriptor is built at each detected corner. The patches are built at every matched points and the similarity between two histograms are modelled by Gaussian distribution. Two set of histograms are built based on RGB and HSV colour space where Neyman-Pearson method decides the best colour model. Diamond search configuration is applied to smooth out the patch position by applying maximum likelihood method. The works by Comaniciu et al. [2] is used as performance comparison. The results show that our algorithm performs better as we have no failure yet lesser average accuracy in tracking fast moving object.

KW - Maximum likelihood

KW - Neyman-Pearson

KW - Patch matching

KW - Tracking observation

UR - http://www.scopus.com/inward/record.url?scp=80052796463&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80052796463&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-23672-3_43

DO - 10.1007/978-3-642-23672-3_43

M3 - Conference contribution

AN - SCOPUS:80052796463

SN - 9783642236716

VL - 6854 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 351

EP - 359

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -