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

研究成果: 書貢獻/報告類型會議貢獻

摘要

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.

原文英語
主出版物標題Digital Human Modeling
主出版物子標題Applications in Health, Safety, Ergonomics, and Risk Management: Health and Safety - 8th International Conference, DHM 2017 Held as Part of HCI International 2017, Proceedings
發行者Springer Verlag
頁面99-107
頁數9
ISBN(列印)9783319584652
DOIs
出版狀態已發佈 - 一月 1 2017
對外發佈Yes
事件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, 加拿大
持續時間: 七月 9 2017七月 14 2017

出版系列

名字Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
10287 LNCS
ISSN(列印)0302-9743
ISSN(電子)1611-3349

其他

其他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
國家加拿大
城市Vancouver
期間7/9/177/14/17

指紋

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

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

引用此文

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. 於 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 (頁 99-107). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 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); 卷 10287 LNCS).

研究成果: 書貢獻/報告類型會議貢獻

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. 於 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), 卷 10287 LNCS, Springer Verlag, 頁 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, 加拿大, 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. 於 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. 頁 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

AN - SCOPUS:85025120700

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 -