Unsupervised Learning from Local Features for Video-based Face Recognition

    Research output: Chapter in Book/Conference paperConference paper

    14 Citations (Scopus)

    Abstract

    This paper presents an unsupervised learning approach to video-based face recognition that does not make any assumptions about the pose, expressions or prior localization of landmarks on the faces. The proposed algorithm exploits spatiotemporal information obtained from local features that are extracted from arbitrary keypoints on faces as opposed to pre-defined landmarks. The algorithm is inherently robust to large scale occlusions as it relies on local features. During unsupervised learning, faces from a video sequence are automatically clustered based on the similarity of their local features and a voting-based algorithm is employed to pick the representative features of each cluster. During recognition, video frames of a probe are sequentially matched to the clusters of all individuals in the gallery and its identity is decided on the basis of best temporally cohesive cluster matches. The proposed algorithms can also detect sudden identity changes in video by utilizing the temporal dimension. The algorithm was tested on the Honda/UCSD video database and a maximum of 99.5% recognition rate was achieved.
    Original languageEnglish
    Title of host publicationProceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008
    EditorsF. Beljaars, P. Stathis
    Place of PublicationLos Alamitos, California, USA
    PublisherIEEE, Institute of Electrical and Electronics Engineers
    Pages1-6
    ISBN (Electronic)9781424421541 , 9781424421534
    Publication statusPublished - Sep 2008
    EventUnsupervised Learning from Local Features for Video-based Face Recognition - Amsterdam, The Netherlands
    Duration: 1 Jan 2008 → …

    Conference

    ConferenceUnsupervised Learning from Local Features for Video-based Face Recognition
    Period1/01/08 → …

    Fingerprint

    Unsupervised learning
    Face recognition

    Cite this

    Mian, A. (2008). Unsupervised Learning from Local Features for Video-based Face Recognition. In F. Beljaars, & P. Stathis (Eds.), Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008 (pp. 1-6). Los Alamitos, California, USA: IEEE, Institute of Electrical and Electronics Engineers.
    Mian, Ajmal. / Unsupervised Learning from Local Features for Video-based Face Recognition. Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008. editor / F. Beljaars ; P. Stathis. Los Alamitos, California, USA : IEEE, Institute of Electrical and Electronics Engineers, 2008. pp. 1-6
    @inproceedings{cf2394f1fde94bb6a9e8f2459346a441,
    title = "Unsupervised Learning from Local Features for Video-based Face Recognition",
    abstract = "This paper presents an unsupervised learning approach to video-based face recognition that does not make any assumptions about the pose, expressions or prior localization of landmarks on the faces. The proposed algorithm exploits spatiotemporal information obtained from local features that are extracted from arbitrary keypoints on faces as opposed to pre-defined landmarks. The algorithm is inherently robust to large scale occlusions as it relies on local features. During unsupervised learning, faces from a video sequence are automatically clustered based on the similarity of their local features and a voting-based algorithm is employed to pick the representative features of each cluster. During recognition, video frames of a probe are sequentially matched to the clusters of all individuals in the gallery and its identity is decided on the basis of best temporally cohesive cluster matches. The proposed algorithms can also detect sudden identity changes in video by utilizing the temporal dimension. The algorithm was tested on the Honda/UCSD video database and a maximum of 99.5{\%} recognition rate was achieved.",
    author = "Ajmal Mian",
    year = "2008",
    month = "9",
    language = "English",
    pages = "1--6",
    editor = "F. Beljaars and P. Stathis",
    booktitle = "Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    address = "United States",

    }

    Mian, A 2008, Unsupervised Learning from Local Features for Video-based Face Recognition. in F Beljaars & P Stathis (eds), Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008. IEEE, Institute of Electrical and Electronics Engineers, Los Alamitos, California, USA, pp. 1-6, Unsupervised Learning from Local Features for Video-based Face Recognition, 1/01/08.

    Unsupervised Learning from Local Features for Video-based Face Recognition. / Mian, Ajmal.

    Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008. ed. / F. Beljaars; P. Stathis. Los Alamitos, California, USA : IEEE, Institute of Electrical and Electronics Engineers, 2008. p. 1-6.

    Research output: Chapter in Book/Conference paperConference paper

    TY - GEN

    T1 - Unsupervised Learning from Local Features for Video-based Face Recognition

    AU - Mian, Ajmal

    PY - 2008/9

    Y1 - 2008/9

    N2 - This paper presents an unsupervised learning approach to video-based face recognition that does not make any assumptions about the pose, expressions or prior localization of landmarks on the faces. The proposed algorithm exploits spatiotemporal information obtained from local features that are extracted from arbitrary keypoints on faces as opposed to pre-defined landmarks. The algorithm is inherently robust to large scale occlusions as it relies on local features. During unsupervised learning, faces from a video sequence are automatically clustered based on the similarity of their local features and a voting-based algorithm is employed to pick the representative features of each cluster. During recognition, video frames of a probe are sequentially matched to the clusters of all individuals in the gallery and its identity is decided on the basis of best temporally cohesive cluster matches. The proposed algorithms can also detect sudden identity changes in video by utilizing the temporal dimension. The algorithm was tested on the Honda/UCSD video database and a maximum of 99.5% recognition rate was achieved.

    AB - This paper presents an unsupervised learning approach to video-based face recognition that does not make any assumptions about the pose, expressions or prior localization of landmarks on the faces. The proposed algorithm exploits spatiotemporal information obtained from local features that are extracted from arbitrary keypoints on faces as opposed to pre-defined landmarks. The algorithm is inherently robust to large scale occlusions as it relies on local features. During unsupervised learning, faces from a video sequence are automatically clustered based on the similarity of their local features and a voting-based algorithm is employed to pick the representative features of each cluster. During recognition, video frames of a probe are sequentially matched to the clusters of all individuals in the gallery and its identity is decided on the basis of best temporally cohesive cluster matches. The proposed algorithms can also detect sudden identity changes in video by utilizing the temporal dimension. The algorithm was tested on the Honda/UCSD video database and a maximum of 99.5% recognition rate was achieved.

    M3 - Conference paper

    SP - 1

    EP - 6

    BT - Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008

    A2 - Beljaars, F.

    A2 - Stathis, P.

    PB - IEEE, Institute of Electrical and Electronics Engineers

    CY - Los Alamitos, California, USA

    ER -

    Mian A. Unsupervised Learning from Local Features for Video-based Face Recognition. In Beljaars F, Stathis P, editors, Proceedings of 8th IEEE International Conference on Automatic Face and Gesture Recognition 2008. Los Alamitos, California, USA: IEEE, Institute of Electrical and Electronics Engineers. 2008. p. 1-6