MURAUER: Mapping Unlabeled Real Data for Label AUstERity

Publikation: Beitrag in einer FachzeitschriftArtikelForschungBegutachtung

Abstract

Data labeling for learning 3D hand pose estimation models is a huge effort. Readily available, accurately labeled synthetic data has the potential to reduce the effort. However, to successfully exploit synthetic data, current state-of-the-art methods still require a large amount of labeled real data. In this work, we remove this requirement by learning to map from the features of real data to the features of synthetic data mainly using a large amount of synthetic and unlabeled real data. We exploit unlabeled data using two auxiliary objectives, which enforce that (i) the mapped representation is pose specific and (ii) at the same time, the distributions of real and synthetic data are aligned. While pose specifity is enforced by a self-supervisory signal requiring that the representation is predictive for the appearance from different views, distributions are aligned by an adversarial term. In this way, we can significantly improve the results of the baseline system, which does not use unlabeled data and outperform many recent approaches already with about 1% of the labeled real data. This presents a step towards faster deployment of learning based hand pose estimation, making it accessible for a larger range of applications.
Originalspracheenglisch
Seitenumfang14
FachzeitschriftarXiv.org e-Print archive
PublikationsstatusVeröffentlicht - 23 Nov 2018

Schlagwörter

    Dies zitieren

    MURAUER : Mapping Unlabeled Real Data for Label AUstERity. / Poier, Georg; Opitz, Michael; Schinagl, David; Bischof, Horst.

    in: arXiv.org e-Print archive, 23.11.2018.

    Publikation: Beitrag in einer FachzeitschriftArtikelForschungBegutachtung

    @article{d8cc50797b2749ab893da2964000fed7,
    title = "MURAUER: Mapping Unlabeled Real Data for Label AUstERity",
    abstract = "Data labeling for learning 3D hand pose estimation models is a huge effort. Readily available, accurately labeled synthetic data has the potential to reduce the effort. However, to successfully exploit synthetic data, current state-of-the-art methods still require a large amount of labeled real data. In this work, we remove this requirement by learning to map from the features of real data to the features of synthetic data mainly using a large amount of synthetic and unlabeled real data. We exploit unlabeled data using two auxiliary objectives, which enforce that (i) the mapped representation is pose specific and (ii) at the same time, the distributions of real and synthetic data are aligned. While pose specifity is enforced by a self-supervisory signal requiring that the representation is predictive for the appearance from different views, distributions are aligned by an adversarial term. In this way, we can significantly improve the results of the baseline system, which does not use unlabeled data and outperform many recent approaches already with about 1{\%} of the labeled real data. This presents a step towards faster deployment of learning based hand pose estimation, making it accessible for a larger range of applications.",
    keywords = "cs.CV, I.2.6; I.2.10; I.4.5; I.4.8; I.4.10; I.5.4",
    author = "Georg Poier and Michael Opitz and David Schinagl and Horst Bischof",
    note = "WACV 2019; Project page at https://poier.github.io/murauer",
    year = "2018",
    month = "11",
    day = "23",
    language = "English",
    journal = "arXiv.org e-Print archive",
    publisher = "Cornell University Library",

    }

    TY - JOUR

    T1 - MURAUER

    T2 - Mapping Unlabeled Real Data for Label AUstERity

    AU - Poier, Georg

    AU - Opitz, Michael

    AU - Schinagl, David

    AU - Bischof, Horst

    N1 - WACV 2019; Project page at https://poier.github.io/murauer

    PY - 2018/11/23

    Y1 - 2018/11/23

    N2 - Data labeling for learning 3D hand pose estimation models is a huge effort. Readily available, accurately labeled synthetic data has the potential to reduce the effort. However, to successfully exploit synthetic data, current state-of-the-art methods still require a large amount of labeled real data. In this work, we remove this requirement by learning to map from the features of real data to the features of synthetic data mainly using a large amount of synthetic and unlabeled real data. We exploit unlabeled data using two auxiliary objectives, which enforce that (i) the mapped representation is pose specific and (ii) at the same time, the distributions of real and synthetic data are aligned. While pose specifity is enforced by a self-supervisory signal requiring that the representation is predictive for the appearance from different views, distributions are aligned by an adversarial term. In this way, we can significantly improve the results of the baseline system, which does not use unlabeled data and outperform many recent approaches already with about 1% of the labeled real data. This presents a step towards faster deployment of learning based hand pose estimation, making it accessible for a larger range of applications.

    AB - Data labeling for learning 3D hand pose estimation models is a huge effort. Readily available, accurately labeled synthetic data has the potential to reduce the effort. However, to successfully exploit synthetic data, current state-of-the-art methods still require a large amount of labeled real data. In this work, we remove this requirement by learning to map from the features of real data to the features of synthetic data mainly using a large amount of synthetic and unlabeled real data. We exploit unlabeled data using two auxiliary objectives, which enforce that (i) the mapped representation is pose specific and (ii) at the same time, the distributions of real and synthetic data are aligned. While pose specifity is enforced by a self-supervisory signal requiring that the representation is predictive for the appearance from different views, distributions are aligned by an adversarial term. In this way, we can significantly improve the results of the baseline system, which does not use unlabeled data and outperform many recent approaches already with about 1% of the labeled real data. This presents a step towards faster deployment of learning based hand pose estimation, making it accessible for a larger range of applications.

    KW - cs.CV

    KW - I.2.6; I.2.10; I.4.5; I.4.8; I.4.10; I.5.4

    M3 - Article

    JO - arXiv.org e-Print archive

    JF - arXiv.org e-Print archive

    ER -