Standard

Automatic Depression Analysis using Dynamic Facial Appearance Descriptor and Dirichlet Process Fisher Encoding. / He, Lang; Jiang, Dongmei; Sahli, Hichem.

In: IEEE Transactions on Multimedia, Vol. 21, No. 6, 8501575, 06.2019, p. 1476-1486.

Research output: Contribution to journalArticleResearchpeer-review

Harvard

APA

Vancouver

Author

He, Lang ; Jiang, Dongmei ; Sahli, Hichem. / Automatic Depression Analysis using Dynamic Facial Appearance Descriptor and Dirichlet Process Fisher Encoding. In: IEEE Transactions on Multimedia. 2019 ; Vol. 21, No. 6. pp. 1476-1486.

BibTeX

@article{725a868fd470494faeb432d1ac38718b,
title = "Automatic Depression Analysis using Dynamic Facial Appearance Descriptor and Dirichlet Process Fisher Encoding",
abstract = "Depression causes mood disorders with noticeable problems in day-to-day activities. Current methods of assessing depression depend almost entirely on clinical interviews or questionnaires. They lack systematic and efficient ways of incorporating behavioral observations that are strong indicators of a psychological disorder. To help clinicians effectively and efficiently diagnose depression severity, automated systems, using objective and quantifiable data for depression assessment, are being developed. This paper presents a framework toward estimating a clinical depression-specific score, namely the Beck Depression Inventory-II (BDI-II) score, based on the analysis of facial expressions features. To extract facial dynamic features, we propose a novel dynamic feature descriptor denoted as median robust local binary patterns from three orthogonal planes (MRLBP-TOP), which can capture both the microstructure and macrostructure of facial appearance and dynamics. To aggregate the MRLBP-TOP over an image sequence, we propose a variant to the Fisher vector (FV) encoding scheme, denoted as the Dirichlet process FV (DPFV). DPFV adopts Dirichlet process Gaussian mixture models (DPGMM) to automatically learn the number of GMM mixtures and model parameters. Experimental results on the AVEC2013 and AVEC2014 depression databases have demonstrated the effectiveness of the proposed method.",
keywords = "Depression, Dirichlet process Fisher vector (DPFV), dynamic feature descriptor, median robust local binary patterns from three orthogonal planes (MRLBP-TOP), nonverbal behaviors",
author = "Lang He and Dongmei Jiang and Hichem Sahli",
year = "2019",
month = "6",
doi = "10.1109/TMM.2018.2877129",
language = "English",
volume = "21",
pages = "1476--1486",
journal = "IEEE Transactions on Multimedia",
issn = "1520-9210",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

RIS

TY - JOUR

T1 - Automatic Depression Analysis using Dynamic Facial Appearance Descriptor and Dirichlet Process Fisher Encoding

AU - He, Lang

AU - Jiang, Dongmei

AU - Sahli, Hichem

PY - 2019/6

Y1 - 2019/6

N2 - Depression causes mood disorders with noticeable problems in day-to-day activities. Current methods of assessing depression depend almost entirely on clinical interviews or questionnaires. They lack systematic and efficient ways of incorporating behavioral observations that are strong indicators of a psychological disorder. To help clinicians effectively and efficiently diagnose depression severity, automated systems, using objective and quantifiable data for depression assessment, are being developed. This paper presents a framework toward estimating a clinical depression-specific score, namely the Beck Depression Inventory-II (BDI-II) score, based on the analysis of facial expressions features. To extract facial dynamic features, we propose a novel dynamic feature descriptor denoted as median robust local binary patterns from three orthogonal planes (MRLBP-TOP), which can capture both the microstructure and macrostructure of facial appearance and dynamics. To aggregate the MRLBP-TOP over an image sequence, we propose a variant to the Fisher vector (FV) encoding scheme, denoted as the Dirichlet process FV (DPFV). DPFV adopts Dirichlet process Gaussian mixture models (DPGMM) to automatically learn the number of GMM mixtures and model parameters. Experimental results on the AVEC2013 and AVEC2014 depression databases have demonstrated the effectiveness of the proposed method.

AB - Depression causes mood disorders with noticeable problems in day-to-day activities. Current methods of assessing depression depend almost entirely on clinical interviews or questionnaires. They lack systematic and efficient ways of incorporating behavioral observations that are strong indicators of a psychological disorder. To help clinicians effectively and efficiently diagnose depression severity, automated systems, using objective and quantifiable data for depression assessment, are being developed. This paper presents a framework toward estimating a clinical depression-specific score, namely the Beck Depression Inventory-II (BDI-II) score, based on the analysis of facial expressions features. To extract facial dynamic features, we propose a novel dynamic feature descriptor denoted as median robust local binary patterns from three orthogonal planes (MRLBP-TOP), which can capture both the microstructure and macrostructure of facial appearance and dynamics. To aggregate the MRLBP-TOP over an image sequence, we propose a variant to the Fisher vector (FV) encoding scheme, denoted as the Dirichlet process FV (DPFV). DPFV adopts Dirichlet process Gaussian mixture models (DPGMM) to automatically learn the number of GMM mixtures and model parameters. Experimental results on the AVEC2013 and AVEC2014 depression databases have demonstrated the effectiveness of the proposed method.

KW - Depression

KW - Dirichlet process Fisher vector (DPFV)

KW - dynamic feature descriptor

KW - median robust local binary patterns from three orthogonal planes (MRLBP-TOP)

KW - nonverbal behaviors

UR - http://www.scopus.com/inward/record.url?scp=85055211751&partnerID=8YFLogxK

U2 - 10.1109/TMM.2018.2877129

DO - 10.1109/TMM.2018.2877129

M3 - Article

VL - 21

SP - 1476

EP - 1486

JO - IEEE Transactions on Multimedia

JF - IEEE Transactions on Multimedia

SN - 1520-9210

IS - 6

M1 - 8501575

ER -

ID: 40029195