Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements

Ming Hsu Wang, Bi Hui Chen, Wen Ko Chiou

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

Abstract

Excessive growth hormone secretion will result in acromegaly affect metabolic function. Patients with acromegaly is 2–4 times greater risk of death than the normal. Early diagnosis is the key follow-up treatment of acromegaly. The clinical diagnosis is based on typical acromegaly the face and body features, endocrine and radiological. However, acromegaly diagnosis is still quite deferred. Typical acromegaly, with the symptoms and appearance, the physician can diagnose. Obvious early symptoms, diagnosis is not easy. As imaging technology advances, one after another to explore the diagnosis of acromegaly, however, did not the size of the stereoscopic 3D image. The aim of this study is to compare the compare the Receiver operating characteristic (ROC) and discriminant analysis for acromegaly detection by three dimensional facial measurements. To explore the difference of detection rate between the two analysis methods. The result shows that the accuracies of three categories from the univariate discriminant analysis, the lateral angles displayed the highest accuracy between all three categories in the female but the lowest rate for the ROC analysis. However, the lateral angles displayed the lowest accuracy between all three categories in the male and the lowest rate for the ROC analysis. The lateral angles, calculated from the two prominent variables, made a larger difference than the other two categories. From the result, it shows that the accuracy difference analysis between the two analysis methods in both genders. The difference could come from the different operation of the analysis methods. It could use the different analysis method to analyze the different facial dimensions for the acromegaly detection in the future and increase the accuracy for disease detection.

Original languageEnglish
Title of host publicationDigital Human Modeling
Subtitle of host publicationApplications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings
PublisherSpringer Verlag
Pages99-107
Number of pages9
ISBN (Print)9783319584652
DOIs
Publication statusPublished - Jan 1 2017
Externally publishedYes
Event8th International Conference on Digital Human Modeling and Applications in Health, Safety, Ergonomics, and Risk Management, DHM 2017, held as part of 19th International Conference on Human-Computer Interaction, HCI 2017 - Vancouver, Canada
Duration: Jul 9 2017Jul 14 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10287 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other8th International Conference on Digital Human Modeling and Applications in Health, Safety, Ergonomics, and Risk Management, DHM 2017, held as part of 19th International Conference on Human-Computer Interaction, HCI 2017
CountryCanada
CityVancouver
Period7/9/177/14/17

Fingerprint

Operating Characteristics
Discriminant analysis
Discriminant Analysis
Receiver
Three-dimensional
Lateral
Lowest
Angle
Growth Hormone
Hormones
Secretion
3D Image
Univariate
High Accuracy
Imaging techniques
Imaging
Face

Keywords

  • Acromegaly
  • Discriminant analysis
  • Receiver operating characteristic

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Wang, M. H., Chen, B. H., & Chiou, W. K. (2017). Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements. In Digital Human Modeling: Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings (pp. 99-107). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10287 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-58466-9_10

Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements. / Wang, Ming Hsu; Chen, Bi Hui; Chiou, Wen Ko.

Digital Human Modeling: Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings. Springer Verlag, 2017. p. 99-107 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10287 LNCS).

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

Wang, MH, Chen, BH & Chiou, WK 2017, Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements. in Digital Human Modeling: Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10287 LNCS, Springer Verlag, pp. 99-107, 8th International Conference on Digital Human Modeling and Applications in Health, Safety, Ergonomics, and Risk Management, DHM 2017, held as part of 19th International Conference on Human-Computer Interaction, HCI 2017, Vancouver, Canada, 7/9/17. https://doi.org/10.1007/978-3-319-58466-9_10
Wang MH, Chen BH, Chiou WK. Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements. In Digital Human Modeling: Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings. Springer Verlag. 2017. p. 99-107. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-58466-9_10
Wang, Ming Hsu ; Chen, Bi Hui ; Chiou, Wen Ko. / Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements. Digital Human Modeling: Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings. Springer Verlag, 2017. pp. 99-107 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d00bf38298e14b4790e027f7452ab985,
title = "Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements",
abstract = "Excessive growth hormone secretion will result in acromegaly affect metabolic function. Patients with acromegaly is 2–4 times greater risk of death than the normal. Early diagnosis is the key follow-up treatment of acromegaly. The clinical diagnosis is based on typical acromegaly the face and body features, endocrine and radiological. However, acromegaly diagnosis is still quite deferred. Typical acromegaly, with the symptoms and appearance, the physician can diagnose. Obvious early symptoms, diagnosis is not easy. As imaging technology advances, one after another to explore the diagnosis of acromegaly, however, did not the size of the stereoscopic 3D image. The aim of this study is to compare the compare the Receiver operating characteristic (ROC) and discriminant analysis for acromegaly detection by three dimensional facial measurements. To explore the difference of detection rate between the two analysis methods. The result shows that the accuracies of three categories from the univariate discriminant analysis, the lateral angles displayed the highest accuracy between all three categories in the female but the lowest rate for the ROC analysis. However, the lateral angles displayed the lowest accuracy between all three categories in the male and the lowest rate for the ROC analysis. The lateral angles, calculated from the two prominent variables, made a larger difference than the other two categories. From the result, it shows that the accuracy difference analysis between the two analysis methods in both genders. The difference could come from the different operation of the analysis methods. It could use the different analysis method to analyze the different facial dimensions for the acromegaly detection in the future and increase the accuracy for disease detection.",
keywords = "Acromegaly, Discriminant analysis, Receiver operating characteristic",
author = "Wang, {Ming Hsu} and Chen, {Bi Hui} and Chiou, {Wen Ko}",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-58466-9_10",
language = "English",
isbn = "9783319584652",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "99--107",
booktitle = "Digital Human Modeling",

}

TY - GEN

T1 - Compare the receiver operating characteristic (ROC) and linear discriminant analysis (LDA) for acromegaly detection by three-dimensional facial measurements

AU - Wang, Ming Hsu

AU - Chen, Bi Hui

AU - Chiou, Wen Ko

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Excessive growth hormone secretion will result in acromegaly affect metabolic function. Patients with acromegaly is 2–4 times greater risk of death than the normal. Early diagnosis is the key follow-up treatment of acromegaly. The clinical diagnosis is based on typical acromegaly the face and body features, endocrine and radiological. However, acromegaly diagnosis is still quite deferred. Typical acromegaly, with the symptoms and appearance, the physician can diagnose. Obvious early symptoms, diagnosis is not easy. As imaging technology advances, one after another to explore the diagnosis of acromegaly, however, did not the size of the stereoscopic 3D image. The aim of this study is to compare the compare the Receiver operating characteristic (ROC) and discriminant analysis for acromegaly detection by three dimensional facial measurements. To explore the difference of detection rate between the two analysis methods. The result shows that the accuracies of three categories from the univariate discriminant analysis, the lateral angles displayed the highest accuracy between all three categories in the female but the lowest rate for the ROC analysis. However, the lateral angles displayed the lowest accuracy between all three categories in the male and the lowest rate for the ROC analysis. The lateral angles, calculated from the two prominent variables, made a larger difference than the other two categories. From the result, it shows that the accuracy difference analysis between the two analysis methods in both genders. The difference could come from the different operation of the analysis methods. It could use the different analysis method to analyze the different facial dimensions for the acromegaly detection in the future and increase the accuracy for disease detection.

AB - Excessive growth hormone secretion will result in acromegaly affect metabolic function. Patients with acromegaly is 2–4 times greater risk of death than the normal. Early diagnosis is the key follow-up treatment of acromegaly. The clinical diagnosis is based on typical acromegaly the face and body features, endocrine and radiological. However, acromegaly diagnosis is still quite deferred. Typical acromegaly, with the symptoms and appearance, the physician can diagnose. Obvious early symptoms, diagnosis is not easy. As imaging technology advances, one after another to explore the diagnosis of acromegaly, however, did not the size of the stereoscopic 3D image. The aim of this study is to compare the compare the Receiver operating characteristic (ROC) and discriminant analysis for acromegaly detection by three dimensional facial measurements. To explore the difference of detection rate between the two analysis methods. The result shows that the accuracies of three categories from the univariate discriminant analysis, the lateral angles displayed the highest accuracy between all three categories in the female but the lowest rate for the ROC analysis. However, the lateral angles displayed the lowest accuracy between all three categories in the male and the lowest rate for the ROC analysis. The lateral angles, calculated from the two prominent variables, made a larger difference than the other two categories. From the result, it shows that the accuracy difference analysis between the two analysis methods in both genders. The difference could come from the different operation of the analysis methods. It could use the different analysis method to analyze the different facial dimensions for the acromegaly detection in the future and increase the accuracy for disease detection.

KW - Acromegaly

KW - Discriminant analysis

KW - Receiver operating characteristic

UR - http://www.scopus.com/inward/record.url?scp=85025120700&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85025120700&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-58466-9_10

DO - 10.1007/978-3-319-58466-9_10

M3 - Conference contribution

SN - 9783319584652

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

SP - 99

EP - 107

BT - Digital Human Modeling

PB - Springer Verlag

ER -