Comparison point tracking with manual tracking in facial palsy analyzer

Park Minjae, Kwang Suk Park

Research output: Contribution to journalConference articleResearchpeer-review

Abstract

Purpose: We developed the asymmetry analyzing system for facial palsy patient using point tracking technique. In last study, we showed initial position of tracking and searching window size is able to affect the tracking performance. However, because reliability of point tracking does not exactly represent overall analyzer performance, analysis of error distance of point tracking from manual tracking is performed. Method: At first, the variation of manual point on the same image should be defined as 10 frame standard deviation with moving window using heading (not-moving) images. Then, error distance between point tracking and manual tracking with 7-pixel and 14-pixel window configuration is evaluated. Result: The standard deviations of first 30 frames are significantly different from next 30 frames by T-test with p-value (2.2961e-006, 0.00075877, 3.3033e-006, 5.5717e-007), and mean standard deviation is 1.5282 (3.0564 for 95.4%). Mean errors over all frame with window 7 is 5.7596, 2.2364, 21.6227, 3.2459 and with window 14 is 8.5959, 2.0236, 4.8618, 3.0331 that are different from window 7 configuration by paired T-test with p-value 0, 0, 0, 0. 265. Conclusion: There is variance of manual tracking and large searching window configuration shows better performance as last study.

Original languageEnglish
Pages (from-to)1626-1628
Number of pages3
JournalIFMBE Proceedings
Volume14
Issue number1
StatePublished - 1 Jan 2007
Event10th World Congress on Medical Physics and Biomedical Engineering, WC 2006 - Seoul, Korea, Republic of
Duration: 27 Aug 20061 Sep 2006

Fingerprint

Pixels

Keywords

  • Error factor analysis
  • Manual tracking
  • Point tracking

Cite this

@article{4652b6af610a44e183276f852f4abcf0,
title = "Comparison point tracking with manual tracking in facial palsy analyzer",
abstract = "Purpose: We developed the asymmetry analyzing system for facial palsy patient using point tracking technique. In last study, we showed initial position of tracking and searching window size is able to affect the tracking performance. However, because reliability of point tracking does not exactly represent overall analyzer performance, analysis of error distance of point tracking from manual tracking is performed. Method: At first, the variation of manual point on the same image should be defined as 10 frame standard deviation with moving window using heading (not-moving) images. Then, error distance between point tracking and manual tracking with 7-pixel and 14-pixel window configuration is evaluated. Result: The standard deviations of first 30 frames are significantly different from next 30 frames by T-test with p-value (2.2961e-006, 0.00075877, 3.3033e-006, 5.5717e-007), and mean standard deviation is 1.5282 (3.0564 for 95.4{\%}). Mean errors over all frame with window 7 is 5.7596, 2.2364, 21.6227, 3.2459 and with window 14 is 8.5959, 2.0236, 4.8618, 3.0331 that are different from window 7 configuration by paired T-test with p-value 0, 0, 0, 0. 265. Conclusion: There is variance of manual tracking and large searching window configuration shows better performance as last study.",
keywords = "Error factor analysis, Manual tracking, Point tracking",
author = "Park Minjae and Park, {Kwang Suk}",
year = "2007",
month = "1",
day = "1",
language = "English",
volume = "14",
pages = "1626--1628",
journal = "IFMBE Proceedings",
issn = "1680-0737",
publisher = "Springer Verlag",
number = "1",

}

Comparison point tracking with manual tracking in facial palsy analyzer. / Minjae, Park; Park, Kwang Suk.

In: IFMBE Proceedings, Vol. 14, No. 1, 01.01.2007, p. 1626-1628.

Research output: Contribution to journalConference articleResearchpeer-review

TY - JOUR

T1 - Comparison point tracking with manual tracking in facial palsy analyzer

AU - Minjae, Park

AU - Park, Kwang Suk

PY - 2007/1/1

Y1 - 2007/1/1

N2 - Purpose: We developed the asymmetry analyzing system for facial palsy patient using point tracking technique. In last study, we showed initial position of tracking and searching window size is able to affect the tracking performance. However, because reliability of point tracking does not exactly represent overall analyzer performance, analysis of error distance of point tracking from manual tracking is performed. Method: At first, the variation of manual point on the same image should be defined as 10 frame standard deviation with moving window using heading (not-moving) images. Then, error distance between point tracking and manual tracking with 7-pixel and 14-pixel window configuration is evaluated. Result: The standard deviations of first 30 frames are significantly different from next 30 frames by T-test with p-value (2.2961e-006, 0.00075877, 3.3033e-006, 5.5717e-007), and mean standard deviation is 1.5282 (3.0564 for 95.4%). Mean errors over all frame with window 7 is 5.7596, 2.2364, 21.6227, 3.2459 and with window 14 is 8.5959, 2.0236, 4.8618, 3.0331 that are different from window 7 configuration by paired T-test with p-value 0, 0, 0, 0. 265. Conclusion: There is variance of manual tracking and large searching window configuration shows better performance as last study.

AB - Purpose: We developed the asymmetry analyzing system for facial palsy patient using point tracking technique. In last study, we showed initial position of tracking and searching window size is able to affect the tracking performance. However, because reliability of point tracking does not exactly represent overall analyzer performance, analysis of error distance of point tracking from manual tracking is performed. Method: At first, the variation of manual point on the same image should be defined as 10 frame standard deviation with moving window using heading (not-moving) images. Then, error distance between point tracking and manual tracking with 7-pixel and 14-pixel window configuration is evaluated. Result: The standard deviations of first 30 frames are significantly different from next 30 frames by T-test with p-value (2.2961e-006, 0.00075877, 3.3033e-006, 5.5717e-007), and mean standard deviation is 1.5282 (3.0564 for 95.4%). Mean errors over all frame with window 7 is 5.7596, 2.2364, 21.6227, 3.2459 and with window 14 is 8.5959, 2.0236, 4.8618, 3.0331 that are different from window 7 configuration by paired T-test with p-value 0, 0, 0, 0. 265. Conclusion: There is variance of manual tracking and large searching window configuration shows better performance as last study.

KW - Error factor analysis

KW - Manual tracking

KW - Point tracking

UR - http://www.scopus.com/inward/record.url?scp=84958268034&partnerID=8YFLogxK

M3 - Conference article

VL - 14

SP - 1626

EP - 1628

JO - IFMBE Proceedings

JF - IFMBE Proceedings

SN - 1680-0737

IS - 1

ER -