Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities

Mohammad Ahsanul Haque, Ruben B. Bautista, Fatemeh Noroozi, Kaustubh Kulkarni, Christian B. Laursen, Ramin Irani, Marco Bellantonio, Sergio Escalera, Gholamreza Anbarjafari, Kamal Nasrollahi, Ole Kæseler Andersen, Erika Geraldina Spaich, Thomas B. Moeslund

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

7 Citationer (Scopus)
459 Downloads (Pure)

Resumé

Pain is a symptom of many disorders associated with actual or potential tissue damage in human body. Managing pain is not only a duty but also highly cost prone. The most primitive state of pain management is the assessment of pain. Traditionally it was accomplished by self-report or visual inspection by experts. However, automatic pain assessment systems from facial videos are also rapidly evolving due to the need of managing pain in a robust and cost effective way. Among different challenges of automatic pain assessment from facial video data two issues are increasingly prevalent: first, exploiting both spatial and temporal information of the face to assess pain level, and second, incorporating multiple visual modalities to capture complementary face information related to pain. Most works in the literature focus on merely exploiting spatial information on chromatic (RGB) video data on shallow learning scenarios. However, employing deep learning techniques for spatio-temporal analysis considering Depth (D) and Thermal (T) along with RGB has high potential in this area. In this paper, we present the first state-of-the-art publicly available database, 'Multimodal Intensity Pain (MIntPAIN)' database, for RGBDT pain level recognition in sequences. We provide a first baseline results including 5 pain levels recognition by analyzing independent visual modalities and their fusion with CNN and LSTM models. From the experimental evaluation we observe that fusion of modalities helps to enhance recognition performance of pain levels in comparison to isolated ones. In particular, the combination of RGB, D, and T in an early fusion fashion achieved the best recognition rate.
OriginalsprogEngelsk
TitelProceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018
Antal sider8
ForlagIEEE
Publikationsdato5 jun. 2018
Sider250-257
ISBN (Elektronisk)978-1-5386-2335-0
DOI
StatusUdgivet - 5 jun. 2018
Begivenhed13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018 - Xi'an, Kina
Varighed: 15 maj 201819 maj 2018

Konference

Konference13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018
LandKina
ByXi'an
Periode15/05/201819/05/2018

Fingerprint

Fusion reactions
Costs
Inspection
Tissue
Hot Temperature
Deep learning

Citer dette

Haque, M. A., B. Bautista, R., Noroozi, F., Kulkarni, K., Laursen, C. B., Irani, R., ... Moeslund, T. B. (2018). Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities. I Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018 (s. 250-257). IEEE. https://doi.org/10.1109/FG.2018.00044
Haque, Mohammad Ahsanul ; B. Bautista, Ruben ; Noroozi, Fatemeh ; Kulkarni, Kaustubh ; Laursen, Christian B. ; Irani, Ramin ; Bellantonio, Marco ; Escalera, Sergio ; Anbarjafari, Gholamreza ; Nasrollahi, Kamal ; Andersen, Ole Kæseler ; Spaich, Erika Geraldina ; Moeslund, Thomas B. / Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities. Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. IEEE, 2018. s. 250-257
@inproceedings{c1ff9f563d3e42438293fd0372b1c119,
title = "Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities",
abstract = "Pain is a symptom of many disorders associated with actual or potential tissue damage in human body. Managing pain is not only a duty but also highly cost prone. The most primitive state of pain management is the assessment of pain. Traditionally it was accomplished by self-report or visual inspection by experts. However, automatic pain assessment systems from facial videos are also rapidly evolving due to the need of managing pain in a robust and cost effective way. Among different challenges of automatic pain assessment from facial video data two issues are increasingly prevalent: first, exploiting both spatial and temporal information of the face to assess pain level, and second, incorporating multiple visual modalities to capture complementary face information related to pain. Most works in the literature focus on merely exploiting spatial information on chromatic (RGB) video data on shallow learning scenarios. However, employing deep learning techniques for spatio-temporal analysis considering Depth (D) and Thermal (T) along with RGB has high potential in this area. In this paper, we present the first state-of-the-art publicly available database, 'Multimodal Intensity Pain (MIntPAIN)' database, for RGBDT pain level recognition in sequences. We provide a first baseline results including 5 pain levels recognition by analyzing independent visual modalities and their fusion with CNN and LSTM models. From the experimental evaluation we observe that fusion of modalities helps to enhance recognition performance of pain levels in comparison to isolated ones. In particular, the combination of RGB, D, and T in an early fusion fashion achieved the best recognition rate.",
keywords = "Database, Deep Learning, Depth, LSTM, Multimodal, Pain, RGB, RGBDT, Thermal, Video, Vision, Visual",
author = "Haque, {Mohammad Ahsanul} and {B. Bautista}, Ruben and Fatemeh Noroozi and Kaustubh Kulkarni and Laursen, {Christian B.} and Ramin Irani and Marco Bellantonio and Sergio Escalera and Gholamreza Anbarjafari and Kamal Nasrollahi and Andersen, {Ole K{\ae}seler} and Spaich, {Erika Geraldina} and Moeslund, {Thomas B.}",
note = "DNRF121",
year = "2018",
month = "6",
day = "5",
doi = "10.1109/FG.2018.00044",
language = "English",
pages = "250--257",
booktitle = "Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018",
publisher = "IEEE",
address = "United States",

}

Haque, MA, B. Bautista, R, Noroozi, F, Kulkarni, K, Laursen, CB, Irani, R, Bellantonio, M, Escalera, S, Anbarjafari, G, Nasrollahi, K, Andersen, OK, Spaich, EG & Moeslund, TB 2018, Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities. i Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. IEEE, s. 250-257, 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018, Xi'an, Kina, 15/05/2018. https://doi.org/10.1109/FG.2018.00044

Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities. / Haque, Mohammad Ahsanul; B. Bautista, Ruben; Noroozi, Fatemeh; Kulkarni, Kaustubh; Laursen, Christian B.; Irani, Ramin; Bellantonio, Marco; Escalera, Sergio; Anbarjafari, Gholamreza; Nasrollahi, Kamal; Andersen, Ole Kæseler; Spaich, Erika Geraldina; Moeslund, Thomas B.

Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. IEEE, 2018. s. 250-257.

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

TY - GEN

T1 - Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities

AU - Haque, Mohammad Ahsanul

AU - B. Bautista, Ruben

AU - Noroozi, Fatemeh

AU - Kulkarni, Kaustubh

AU - Laursen, Christian B.

AU - Irani, Ramin

AU - Bellantonio, Marco

AU - Escalera, Sergio

AU - Anbarjafari, Gholamreza

AU - Nasrollahi, Kamal

AU - Andersen, Ole Kæseler

AU - Spaich, Erika Geraldina

AU - Moeslund, Thomas B.

N1 - DNRF121

PY - 2018/6/5

Y1 - 2018/6/5

N2 - Pain is a symptom of many disorders associated with actual or potential tissue damage in human body. Managing pain is not only a duty but also highly cost prone. The most primitive state of pain management is the assessment of pain. Traditionally it was accomplished by self-report or visual inspection by experts. However, automatic pain assessment systems from facial videos are also rapidly evolving due to the need of managing pain in a robust and cost effective way. Among different challenges of automatic pain assessment from facial video data two issues are increasingly prevalent: first, exploiting both spatial and temporal information of the face to assess pain level, and second, incorporating multiple visual modalities to capture complementary face information related to pain. Most works in the literature focus on merely exploiting spatial information on chromatic (RGB) video data on shallow learning scenarios. However, employing deep learning techniques for spatio-temporal analysis considering Depth (D) and Thermal (T) along with RGB has high potential in this area. In this paper, we present the first state-of-the-art publicly available database, 'Multimodal Intensity Pain (MIntPAIN)' database, for RGBDT pain level recognition in sequences. We provide a first baseline results including 5 pain levels recognition by analyzing independent visual modalities and their fusion with CNN and LSTM models. From the experimental evaluation we observe that fusion of modalities helps to enhance recognition performance of pain levels in comparison to isolated ones. In particular, the combination of RGB, D, and T in an early fusion fashion achieved the best recognition rate.

AB - Pain is a symptom of many disorders associated with actual or potential tissue damage in human body. Managing pain is not only a duty but also highly cost prone. The most primitive state of pain management is the assessment of pain. Traditionally it was accomplished by self-report or visual inspection by experts. However, automatic pain assessment systems from facial videos are also rapidly evolving due to the need of managing pain in a robust and cost effective way. Among different challenges of automatic pain assessment from facial video data two issues are increasingly prevalent: first, exploiting both spatial and temporal information of the face to assess pain level, and second, incorporating multiple visual modalities to capture complementary face information related to pain. Most works in the literature focus on merely exploiting spatial information on chromatic (RGB) video data on shallow learning scenarios. However, employing deep learning techniques for spatio-temporal analysis considering Depth (D) and Thermal (T) along with RGB has high potential in this area. In this paper, we present the first state-of-the-art publicly available database, 'Multimodal Intensity Pain (MIntPAIN)' database, for RGBDT pain level recognition in sequences. We provide a first baseline results including 5 pain levels recognition by analyzing independent visual modalities and their fusion with CNN and LSTM models. From the experimental evaluation we observe that fusion of modalities helps to enhance recognition performance of pain levels in comparison to isolated ones. In particular, the combination of RGB, D, and T in an early fusion fashion achieved the best recognition rate.

KW - Database

KW - Deep Learning

KW - Depth

KW - LSTM

KW - Multimodal

KW - Pain

KW - RGB

KW - RGBDT

KW - Thermal

KW - Video

KW - Vision

KW - Visual

UR - http://www.scopus.com/inward/record.url?scp=85049408742&partnerID=8YFLogxK

U2 - 10.1109/FG.2018.00044

DO - 10.1109/FG.2018.00044

M3 - Article in proceeding

SP - 250

EP - 257

BT - Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018

PB - IEEE

ER -

Haque MA, B. Bautista R, Noroozi F, Kulkarni K, Laursen CB, Irani R et al. Deep Multimodal Pain Recognition: A Database and Comparison of Spatio-Temporal Visual Modalities. I Proceedings - 13th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2018. IEEE. 2018. s. 250-257 https://doi.org/10.1109/FG.2018.00044