Evaluating AAM fitting methods for facial expression recognition

Akshay Asthana, Jason Saragih, Michael Wagner, Roland Goecke

    Research output: A Conference proceeding or a Chapter in BookConference contribution

    44 Citations (Scopus)

    Abstract

    The human face is a rich source of information for the viewer and facial expressions are a major component in judging a person's affective state, intention and personality. Facial expressions are an important part of human-human interaction and have the potential to play an equally important part in human-computer interaction. This paper evaluates various active appearance model (AAM) fitting methods, including both the original formulation as well as several state-of-the-art methods, for the task of automatic facial expression recognition. The AAM is a powerful statistical model for modelling and registering deformable objects. The results of the fitting process are used in a facial expression recognition task using a region-based intermediate representation related to action units, with the expression classification task realised using a support vector machine. Experiments are performed for both person-dependent and person-independent setups. Overall, the best facial expression recognition results were obtained by using the iterative error bound minimisation method, which consistently resulted in accurate face model alignment and facial expression recognition even when the initial face detection used to initialise the fitting procedure was poor.
    Original languageEnglish
    Title of host publication3rd International Conference on Affective Computing and Intelligent Interaction and Workshops
    EditorsJeffrey Cohn, Anton Nijholt, Maja Pantic
    Place of PublicationThe Netherlands
    PublisherIEEE, Institute of Electrical and Electronics Engineers
    Pages1-8
    Number of pages8
    Volume1
    ISBN (Print)9781424447992
    DOIs
    Publication statusPublished - 2009
    EventACII 2009 - Amsterdam, Netherlands
    Duration: 10 Sep 200912 Sep 2009

    Conference

    ConferenceACII 2009
    CountryNetherlands
    CityAmsterdam
    Period10/09/0912/09/09

    Fingerprint

    Human computer interaction
    Face recognition
    Support vector machines
    Experiments
    Statistical Models

    Cite this

    Asthana, A., Saragih, J., Wagner, M., & Goecke, R. (2009). Evaluating AAM fitting methods for facial expression recognition. In J. Cohn, A. Nijholt, & M. Pantic (Eds.), 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops (Vol. 1, pp. 1-8). The Netherlands: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ACII.2009.5349489
    Asthana, Akshay ; Saragih, Jason ; Wagner, Michael ; Goecke, Roland. / Evaluating AAM fitting methods for facial expression recognition. 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops. editor / Jeffrey Cohn ; Anton Nijholt ; Maja Pantic. Vol. 1 The Netherlands : IEEE, Institute of Electrical and Electronics Engineers, 2009. pp. 1-8
    @inproceedings{99148343a8db46eb886b3670146e51ab,
    title = "Evaluating AAM fitting methods for facial expression recognition",
    abstract = "The human face is a rich source of information for the viewer and facial expressions are a major component in judging a person's affective state, intention and personality. Facial expressions are an important part of human-human interaction and have the potential to play an equally important part in human-computer interaction. This paper evaluates various active appearance model (AAM) fitting methods, including both the original formulation as well as several state-of-the-art methods, for the task of automatic facial expression recognition. The AAM is a powerful statistical model for modelling and registering deformable objects. The results of the fitting process are used in a facial expression recognition task using a region-based intermediate representation related to action units, with the expression classification task realised using a support vector machine. Experiments are performed for both person-dependent and person-independent setups. Overall, the best facial expression recognition results were obtained by using the iterative error bound minimisation method, which consistently resulted in accurate face model alignment and facial expression recognition even when the initial face detection used to initialise the fitting procedure was poor.",
    author = "Akshay Asthana and Jason Saragih and Michael Wagner and Roland Goecke",
    year = "2009",
    doi = "10.1109/ACII.2009.5349489",
    language = "English",
    isbn = "9781424447992",
    volume = "1",
    pages = "1--8",
    editor = "Jeffrey Cohn and Anton Nijholt and Maja Pantic",
    booktitle = "3rd International Conference on Affective Computing and Intelligent Interaction and Workshops",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    address = "United States",

    }

    Asthana, A, Saragih, J, Wagner, M & Goecke, R 2009, Evaluating AAM fitting methods for facial expression recognition. in J Cohn, A Nijholt & M Pantic (eds), 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops. vol. 1, IEEE, Institute of Electrical and Electronics Engineers, The Netherlands, pp. 1-8, ACII 2009, Amsterdam, Netherlands, 10/09/09. https://doi.org/10.1109/ACII.2009.5349489

    Evaluating AAM fitting methods for facial expression recognition. / Asthana, Akshay; Saragih, Jason; Wagner, Michael; Goecke, Roland.

    3rd International Conference on Affective Computing and Intelligent Interaction and Workshops. ed. / Jeffrey Cohn; Anton Nijholt; Maja Pantic. Vol. 1 The Netherlands : IEEE, Institute of Electrical and Electronics Engineers, 2009. p. 1-8.

    Research output: A Conference proceeding or a Chapter in BookConference contribution

    TY - GEN

    T1 - Evaluating AAM fitting methods for facial expression recognition

    AU - Asthana, Akshay

    AU - Saragih, Jason

    AU - Wagner, Michael

    AU - Goecke, Roland

    PY - 2009

    Y1 - 2009

    N2 - The human face is a rich source of information for the viewer and facial expressions are a major component in judging a person's affective state, intention and personality. Facial expressions are an important part of human-human interaction and have the potential to play an equally important part in human-computer interaction. This paper evaluates various active appearance model (AAM) fitting methods, including both the original formulation as well as several state-of-the-art methods, for the task of automatic facial expression recognition. The AAM is a powerful statistical model for modelling and registering deformable objects. The results of the fitting process are used in a facial expression recognition task using a region-based intermediate representation related to action units, with the expression classification task realised using a support vector machine. Experiments are performed for both person-dependent and person-independent setups. Overall, the best facial expression recognition results were obtained by using the iterative error bound minimisation method, which consistently resulted in accurate face model alignment and facial expression recognition even when the initial face detection used to initialise the fitting procedure was poor.

    AB - The human face is a rich source of information for the viewer and facial expressions are a major component in judging a person's affective state, intention and personality. Facial expressions are an important part of human-human interaction and have the potential to play an equally important part in human-computer interaction. This paper evaluates various active appearance model (AAM) fitting methods, including both the original formulation as well as several state-of-the-art methods, for the task of automatic facial expression recognition. The AAM is a powerful statistical model for modelling and registering deformable objects. The results of the fitting process are used in a facial expression recognition task using a region-based intermediate representation related to action units, with the expression classification task realised using a support vector machine. Experiments are performed for both person-dependent and person-independent setups. Overall, the best facial expression recognition results were obtained by using the iterative error bound minimisation method, which consistently resulted in accurate face model alignment and facial expression recognition even when the initial face detection used to initialise the fitting procedure was poor.

    U2 - 10.1109/ACII.2009.5349489

    DO - 10.1109/ACII.2009.5349489

    M3 - Conference contribution

    SN - 9781424447992

    VL - 1

    SP - 1

    EP - 8

    BT - 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops

    A2 - Cohn, Jeffrey

    A2 - Nijholt, Anton

    A2 - Pantic, Maja

    PB - IEEE, Institute of Electrical and Electronics Engineers

    CY - The Netherlands

    ER -

    Asthana A, Saragih J, Wagner M, Goecke R. Evaluating AAM fitting methods for facial expression recognition. In Cohn J, Nijholt A, Pantic M, editors, 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops. Vol. 1. The Netherlands: IEEE, Institute of Electrical and Electronics Engineers. 2009. p. 1-8 https://doi.org/10.1109/ACII.2009.5349489