Detecting head movements in video-recorded dyadic conversations

Patrizia Paggio; Bart Jongejan; Manex Agirrezabal; Costanza Navarretta

doi:10.1145/3281151.3281152

Detecting head movements in video-recorded dyadic conversations

Patrizia Paggio, Bart Jongejan, Manex Agirrezabal, Costanza Navarretta

Department of Nordic Studies and Linguistics

Abstract

This paper is about the automatic recognition of head movements in videos of face-to-face dyadic conversations. We present an approach where recognition of head movements is casted as a multimodal frame classification problem based on visual and acoustic features. The visual features include velocity, acceleration, and jerk values associated with head movements, while the acoustic ones are pitch and intensity measurements from the co-occuring speech. We present the results obtained by training and testing a number of classifiers on manually annotated data from two conversations. The best performing classifier, a Multilayer Perceptron trained using all the features, obtains 0.75 accuracy and outperforms the mono-modal baseline classifier.

Original language	English
Title of host publication	Proceedings of the International Conference on Multimodal Interaction: Adjunct
Number of pages	6
Place of Publication	New York
Publisher	Association for Computing Machinery
Publication date	16 Oct 2018
Pages	1-6
ISBN (Print)	978-1-4503-6002-9
DOIs	https://doi.org/10.1145/3281151.3281152
Publication status	Published - 16 Oct 2018

Access to Document

10.1145/3281151.3281152

Cite this

Detecting head movements in video-recorded dyadic conversations. / Paggio, Patrizia ; Jongejan, Bart ; Agirrezabal, Manex et al.

Proceedings of the International Conference on Multimodal Interaction: Adjunct. New York : Association for Computing Machinery, 2018. p. 1-6.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

@inproceedings{f810685cd29b4210895ccb8b4b4e2605,

title = "Detecting head movements in video-recorded dyadic conversations",

abstract = "This paper is about the automatic recognition of head movements in videos of face-to-face dyadic conversations. We present an approach where recognition of head movements is casted as a multimodal frame classification problem based on visual and acoustic features. The visual features include velocity, acceleration, and jerk values associated with head movements, while the acoustic ones are pitch and intensity measurements from the co-occuring speech. We present the results obtained by training and testing a number of classifiers on manually annotated data from two conversations. The best performing classifier, a Multilayer Perceptron trained using all the features, obtains 0.75 accuracy and outperforms the mono-modal baseline classifier.",

author = "Patrizia Paggio and Bart Jongejan and Manex Agirrezabal and Costanza Navarretta",

year = "2018",

month = oct,

day = "16",

doi = "10.1145/3281151.3281152",

language = "English",

isbn = "978-1-4503-6002-9",

pages = "1--6",

booktitle = "Proceedings of the International Conference on Multimodal Interaction: Adjunct",

publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Detecting head movements in video-recorded dyadic conversations

AU - Paggio, Patrizia

AU - Jongejan, Bart

AU - Agirrezabal, Manex

AU - Navarretta, Costanza

PY - 2018/10/16

Y1 - 2018/10/16

N2 - This paper is about the automatic recognition of head movements in videos of face-to-face dyadic conversations. We present an approach where recognition of head movements is casted as a multimodal frame classification problem based on visual and acoustic features. The visual features include velocity, acceleration, and jerk values associated with head movements, while the acoustic ones are pitch and intensity measurements from the co-occuring speech. We present the results obtained by training and testing a number of classifiers on manually annotated data from two conversations. The best performing classifier, a Multilayer Perceptron trained using all the features, obtains 0.75 accuracy and outperforms the mono-modal baseline classifier.

AB - This paper is about the automatic recognition of head movements in videos of face-to-face dyadic conversations. We present an approach where recognition of head movements is casted as a multimodal frame classification problem based on visual and acoustic features. The visual features include velocity, acceleration, and jerk values associated with head movements, while the acoustic ones are pitch and intensity measurements from the co-occuring speech. We present the results obtained by training and testing a number of classifiers on manually annotated data from two conversations. The best performing classifier, a Multilayer Perceptron trained using all the features, obtains 0.75 accuracy and outperforms the mono-modal baseline classifier.

UR - https://dl.acm.org/citation.cfm?doid=3281151.3281152

U2 - 10.1145/3281151.3281152

DO - 10.1145/3281151.3281152

M3 - Article in proceedings

SN - 978-1-4503-6002-9

SP - 1

EP - 6

BT - Proceedings of the International Conference on Multimodal Interaction: Adjunct

PB - Association for Computing Machinery

CY - New York

ER -

Detecting head movements in video-recorded dyadic conversations

Abstract

Access to Document

Other files and links

Fingerprint

Cite this