Using Kinect for face recognition under varying poses, expressions, illumination and disguise

B.Y.L. Li, Ajmal Mian, W. Liu, A. Krishna

    Research output: Chapter in Book/Conference paperConference paper

    151 Citations (Scopus)
    952 Downloads (Pure)

    Abstract

    We present an algorithm that uses a low resolution 3D sensor for robust face recognition under challenging conditions. A preprocessing algorithm is proposed which exploits the facial symmetry at the 3D point cloud level to obtain a canonical frontal view, shape and texture, of the faces irrespective of their initial pose. This algorithm also fills holes and smooths the noisy depth data produced by the low resolution sensor. The canonical depth map and texture of a query face are then sparse approximated from separate dictionaries learned from training data. The texture is transformed from the RGB to Discriminant Color Space before sparse coding and the reconstruction errors from the two sparse coding steps are added for individual identities in the dictionary. The query face is assigned the identity with the smallest reconstruction error. Experiments are performed using a publicly available database containing over 5000 facial images (RGB-D) with varying poses, expressions, illumination and disguise, acquired using the Kinect sensor. Recognition rates are 96.7% for the RGB-D data and 88.7% for the noisy depth data alone. Our results justify the feasibility of low resolution 3D sensors for robust face recognition.
    Original languageEnglish
    Title of host publication2013 IEEE Workshop on Applications of Computer Vision (WACV)
    Place of PublicationUnited States
    PublisherIEEE, Institute of Electrical and Electronics Engineers
    Pages186-192
    ISBN (Print)9781467350532
    DOIs
    Publication statusPublished - 2013
    Event2013 IEEE Workshop on Applications of Computer Vision - Tampa, United States
    Duration: 15 Jan 201317 Jan 2013

    Conference

    Conference2013 IEEE Workshop on Applications of Computer Vision
    Abbreviated titleWACV 2013
    CountryUnited States
    CityTampa
    Period15/01/1317/01/13

    Fingerprint

    Face recognition
    Lighting
    Textures
    Sensors
    Glossaries
    Color
    Experiments

    Cite this

    Li, B. Y. L., Mian, A., Liu, W., & Krishna, A. (2013). Using Kinect for face recognition under varying poses, expressions, illumination and disguise. In 2013 IEEE Workshop on Applications of Computer Vision (WACV) (pp. 186-192). United States: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/WACV.2013.6475017
    Li, B.Y.L. ; Mian, Ajmal ; Liu, W. ; Krishna, A. / Using Kinect for face recognition under varying poses, expressions, illumination and disguise. 2013 IEEE Workshop on Applications of Computer Vision (WACV). United States : IEEE, Institute of Electrical and Electronics Engineers, 2013. pp. 186-192
    @inproceedings{8f4775904fa94c91a78f638cca20f064,
    title = "Using Kinect for face recognition under varying poses, expressions, illumination and disguise",
    abstract = "We present an algorithm that uses a low resolution 3D sensor for robust face recognition under challenging conditions. A preprocessing algorithm is proposed which exploits the facial symmetry at the 3D point cloud level to obtain a canonical frontal view, shape and texture, of the faces irrespective of their initial pose. This algorithm also fills holes and smooths the noisy depth data produced by the low resolution sensor. The canonical depth map and texture of a query face are then sparse approximated from separate dictionaries learned from training data. The texture is transformed from the RGB to Discriminant Color Space before sparse coding and the reconstruction errors from the two sparse coding steps are added for individual identities in the dictionary. The query face is assigned the identity with the smallest reconstruction error. Experiments are performed using a publicly available database containing over 5000 facial images (RGB-D) with varying poses, expressions, illumination and disguise, acquired using the Kinect sensor. Recognition rates are 96.7{\%} for the RGB-D data and 88.7{\%} for the noisy depth data alone. Our results justify the feasibility of low resolution 3D sensors for robust face recognition.",
    author = "B.Y.L. Li and Ajmal Mian and W. Liu and A. Krishna",
    year = "2013",
    doi = "10.1109/WACV.2013.6475017",
    language = "English",
    isbn = "9781467350532",
    pages = "186--192",
    booktitle = "2013 IEEE Workshop on Applications of Computer Vision (WACV)",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    address = "United States",

    }

    Li, BYL, Mian, A, Liu, W & Krishna, A 2013, Using Kinect for face recognition under varying poses, expressions, illumination and disguise. in 2013 IEEE Workshop on Applications of Computer Vision (WACV). IEEE, Institute of Electrical and Electronics Engineers, United States, pp. 186-192, 2013 IEEE Workshop on Applications of Computer Vision , Tampa, United States, 15/01/13. https://doi.org/10.1109/WACV.2013.6475017

    Using Kinect for face recognition under varying poses, expressions, illumination and disguise. / Li, B.Y.L.; Mian, Ajmal; Liu, W.; Krishna, A.

    2013 IEEE Workshop on Applications of Computer Vision (WACV). United States : IEEE, Institute of Electrical and Electronics Engineers, 2013. p. 186-192.

    Research output: Chapter in Book/Conference paperConference paper

    TY - GEN

    T1 - Using Kinect for face recognition under varying poses, expressions, illumination and disguise

    AU - Li, B.Y.L.

    AU - Mian, Ajmal

    AU - Liu, W.

    AU - Krishna, A.

    PY - 2013

    Y1 - 2013

    N2 - We present an algorithm that uses a low resolution 3D sensor for robust face recognition under challenging conditions. A preprocessing algorithm is proposed which exploits the facial symmetry at the 3D point cloud level to obtain a canonical frontal view, shape and texture, of the faces irrespective of their initial pose. This algorithm also fills holes and smooths the noisy depth data produced by the low resolution sensor. The canonical depth map and texture of a query face are then sparse approximated from separate dictionaries learned from training data. The texture is transformed from the RGB to Discriminant Color Space before sparse coding and the reconstruction errors from the two sparse coding steps are added for individual identities in the dictionary. The query face is assigned the identity with the smallest reconstruction error. Experiments are performed using a publicly available database containing over 5000 facial images (RGB-D) with varying poses, expressions, illumination and disguise, acquired using the Kinect sensor. Recognition rates are 96.7% for the RGB-D data and 88.7% for the noisy depth data alone. Our results justify the feasibility of low resolution 3D sensors for robust face recognition.

    AB - We present an algorithm that uses a low resolution 3D sensor for robust face recognition under challenging conditions. A preprocessing algorithm is proposed which exploits the facial symmetry at the 3D point cloud level to obtain a canonical frontal view, shape and texture, of the faces irrespective of their initial pose. This algorithm also fills holes and smooths the noisy depth data produced by the low resolution sensor. The canonical depth map and texture of a query face are then sparse approximated from separate dictionaries learned from training data. The texture is transformed from the RGB to Discriminant Color Space before sparse coding and the reconstruction errors from the two sparse coding steps are added for individual identities in the dictionary. The query face is assigned the identity with the smallest reconstruction error. Experiments are performed using a publicly available database containing over 5000 facial images (RGB-D) with varying poses, expressions, illumination and disguise, acquired using the Kinect sensor. Recognition rates are 96.7% for the RGB-D data and 88.7% for the noisy depth data alone. Our results justify the feasibility of low resolution 3D sensors for robust face recognition.

    U2 - 10.1109/WACV.2013.6475017

    DO - 10.1109/WACV.2013.6475017

    M3 - Conference paper

    SN - 9781467350532

    SP - 186

    EP - 192

    BT - 2013 IEEE Workshop on Applications of Computer Vision (WACV)

    PB - IEEE, Institute of Electrical and Electronics Engineers

    CY - United States

    ER -

    Li BYL, Mian A, Liu W, Krishna A. Using Kinect for face recognition under varying poses, expressions, illumination and disguise. In 2013 IEEE Workshop on Applications of Computer Vision (WACV). United States: IEEE, Institute of Electrical and Electronics Engineers. 2013. p. 186-192 https://doi.org/10.1109/WACV.2013.6475017